Minggu, 03 Februari 2013

jurnal perilaku konsumen “sikap dan intensi konsumen” ( internasional )

Consumer attitude toward virtual stores and its correlates

Pengarang       : B.C.Y. Lee
Diterbitkan di  : sciencedirect.com

Abstract
Supported by the growth of information technology, the virtual stores have increased their importance in the modern marketing environment. The purpose of this research is to investigate the relationships between consumer attitude toward virtual stores and its correlates. A conceptual model with four postulated hypotheses is proposed and verified by empirical data. Our study results show that consumer risk averseness is negatively related to consumer attitude, whereas consumer convenience orientation and the impulse tendency are both positively related with consumer attitude toward virtual stores. Implications for practitioners and suggestions for further research are also provided.

1. Introduction

The rapid growth of information technology has enhanced the importance of virtual stores as a marketing channel (Burke, 2002; Chen and Tan, 2004; Jarvenpaa and Todd, 1997; Korgaonkar, 1984; Vrechopoulos et al., 2004). The “virtual store” represents a private retailer, without a fixed showroom and face-to-face contact, utilizing information techniques and the media to communicate with consumers and pursue marketing goals. Virtual stores include catalog shopping, TV shopping channels, Internet stores, etc. Of these, the Internet is perhaps the most visible innovation, attracting considerable media and commercial attention (Burke, 2002; Chen and Tan, 2004). In the USA, 50% of households have one or more computers, and 42% include a family member who uses the Internet at home (Burke, 2002). About 42% of the major suppliers in a variety of industries have begun to sell directly to consumers over the Internet (Burke, 2002). In addition to the Internet, TV shopping and catalog mailing have also grown. Each year, over 12 billion direct mail catalogs are mailed in the United States alone; and TV shopping channels have improved their product categories and sales revenue (Basu et al., 1995).

Without a physical and fixed outlet to show and promote their products, many virtual stores employ direct marketing skills to attract and communicate with potential consumers (Mehta and Sivadas, 1995). Direct marketing is an interactive system of marketing that uses one or more advertising media to affect a measurable response and/or transaction at any location, with this activity stored on a database (Stone, 1997). Direct marketing is growing at twice the rate of traditional retailing methods (May, 1989). In fact, more money is spent on direct marketing programs and solicitations than on magazine or television advertising (Akaah et al., 1995). Sometimes direct marketing has only been used to provide information and to support sales in traditional channels (Chiang et al., 2003).

Since the virtual stores have many advantages, including lower operational costs, 24-h service all year round, greater product diversity, and the ability to reach distant customers (Tapp, 2001; Tiwana, 1998), they have changed the ways of marketing communication, promotion, and exchange. From the consumer’s perspective, he/she must face a new way of selecting and purchasing products. According to the theory of reasoned action (Ajzen and Fishbein, 1980), consumer attitude affects consumer behavioral intention, thus influencing consumer purchasing behavior. To maximize the utility of virtual stores, marketers need to understand consumer attitude toward virtual stores and its correlates. Although issues related to virtual stores have been examined in the literature (Akaah et al., 1995; Burke, 2002; Chen and Tan, 2004; Dirk and Leunis, 1999; Liang and Huang, 1998; Vrechopoulos et al., 2004), consumer attitude toward virtual stores and its antecedents have received little attention.

The purpose of this research is to investigate consumer attitude toward virtual stores and its correlates. Focus group discussion is used to explore the potential antecedents of consumer attitude toward virtual stores. A conceptual model with four hypotheses is proposed and empirically examined by field data. Structural equation modeling with the LISREL program is used to estimate the structural coefficients and test the hypotheses. Implications for managers of virtual stores and suggestions for further research are also provided.

The purpose of this research is to investigate consumer attitude toward virtual stores and its correlates. Focus group discussion is used to explore the potential antecedents of consumer attitude toward virtual stores. A conceptual model with four hypotheses is proposed and empirically examined by field data. Structural equation modeling with the LISREL program is used to estimate the structural coefficients and test the hypotheses. Implications for managers of virtual stores and suggestions for further research are also provided.

2. Hypotheses and proposed model

Since new information technology is changing the types and business models of virtual stores, this article focuses on consumer attitude toward virtual stores in general, rather than any specific store or specific product category. This work is thus consistent with that on the construction of measures that evaluate general consumer attitudes or tendencies, such as attitude toward advertising in general rather than toward a particular advertisement.

2.1. Consumer attitude toward virtual store and its antecedents

“Attitude” denotes a learned predisposition to respond to an object in a consistently favorable or unfavorable way. The fact that attitudes are learned means that they will be affected by information and experiences (Wilkie, 1994). Since attitudes cannot be observed directly, they are mental positions that marketers must try to infer through research measures. The fact that attitudes are predispositions to respond indicates their relationship with actual consumer behavior.

To explore the possible antecedents of consumer attitude toward virtual stores, a focus group discussion (FGD) technique was conducted. Ten undergraduate students in business school were invited to participate in the discussion. Various factors that relate to virtual stores were raised and discussed, including risk, efficiency, impulsiveness, product type etc. From the consumer point of view, lack of personal contact implies that they must carry a higher purchasing risk, but they can save travel time and shopping effort. In addition, the influence of the consumer’s level of impulsiveness has been mentioned frequently. As a result, risk averseness, convenience orientation, and impulse tendency were selected as possible antecedents of consumer attitude toward virtual stores. We discuss these factors and their possible influence on consumer attitude in turn.

2.1.1. Risk averseness

“Risk” in relation to choosing brands is the probability of occurrence of a problem with a particular product of a certain brand multiplied by the negative consequences of that problem (Peter and Ryan, 1976). Risk can be conceptualized as an objective characteristic of a given situation, but the assessment of risk involves an individual bringing his or her own characteristics to the situation and also the accurate appraisal of risk (Conchar et al., 2004). Following from subjective expected utility theory, risk is modeled by reflecting the decision-maker’s response to uncertain outcomes defined in terms of specific probabilities of risk (Mitchell, 1999). When a consumer makes a purchase decision, “risk” implies “greater consequences of making a mistake” and “degree of inconvenience of making a mistake” (Batra and Sinha, 2000). Havlena and DeSarbo (1991) described the multidimensional nature of perceived consumer risk, involving performance, financial, safety, social, psychological, and time/opportunity dimensions. The perceived risk can powerfully influence consumer behavior. Mitchell (1992) argued that perceived risk influences the five stages of the consumer decision process, which are problem recognition, pre-purchase information search, evaluation of alternatives, purchase decision and post-purchase behavior.

Consumers’ perceived risk in virtual stores has received considerable attention in recent years (Bhatnagar and Ghose, 2004; Chen and He, 2003; Forsythe and Shi, 2003; Gupta et al., 2004; Lim, 2003; Milne and Culnan, 2004; Miyazaki and Fernandez, 2001; Pavlou, 2003). Pavlou (2003) stated that the e-commerce environment creates both economic and privacy risks for consumers, and their perceived risk was strongly related to their intention to transact in an e-commerce environment. Through focus group discussions, Lim (2003) found that Internet consumers perceive three sources of risk in B2C e-commerce: technology, vendor, and product. Milne and Culnan (2004) suggested that privacy notices are an important means of reducing the risk of the second exchange by providing consumers with information concerning the organization’s information practices. In addition to Internet shopping, studies have consistently shown that consumers perceive higher risks in non-store shopping formats, such as telephone shopping, mail order, catalog, direct sales, and catalog showroom (Bhatnagar and Ghose, 2004; Peterson et al., 1989).

Risk aversion is defined as “the extent to which people feel threatened by ambiguous situations, and have created beliefs and institutions that try to avoid these” (Hofstede and Bond, 1984). People with higher risk aversion are inclined to feel threatened by risky and ambiguous situations (Hofstede, 1991). In addition, after making purchases through one channel, either electronic or traditional, risk-averse consumers tend to be more loyal customers than risk-neutral consumers (Gupta et al., 2004).

Consumers will face various risks when purchasing in virtual stores. First, without direct investigation on a product before making a purchase decision, consumers must carry a higher product performance risk. Secondly, the new exchange process in the virtual stores will increase consumers’ financial risk (i.e., payment before delivery or payment by credit card). Finally, when lacking the personal explanation of product knowledge, consumers increase their probability of inappropriate product use, which heightens the safety risks. Given the potential risks in virtual stores, consumers’ risk averseness should negatively affect their attitudes toward virtual stores.


2.1.2. Convenience orientation

Convenience is considered to encompass various utilities, including the time, place, acquisition and use of a product or service. The construct of convenience has two main dimensions: time and energy (McEnally and Brown, 1998). The time dimension usually refers to lack of time or to time pressure, either of which leads to convenience-oriented behavior; while the energy is subdivided into mental energy, usually in the form of the effort involved in planning ahead, and physical energy, which involves doing something to obtain a desired product or service (Marquis, 2005). To a degree, convenience-oriented consumption represents a point of convergence between the coincidentally increasing affluence and time-consciousness of the contemporary consumer. Convenience-oriented consumption is distinguished by the fact that it is prompted by the dual motives of (1) satisfying some immediate want or need, and (2) releasing time and/or energy for alternative uses (Anderson, 1972).

Convenience-oriented consumers do differ from cost-oriented consumers in terms of demographic, lifestyle, and product value. There is growing recognition that consumers are constrained by two budgets: a “money” budget and a “time” budget (Engel and Blackwell, 1982). Those consumers that are the most time-constrained are likely to value products that offer convenience features, whereas money-constrained consumers are more likely to place emphasis on product costs. The convenience-oriented consumer is defined as one who seeks to “accomplish a task in the shortest time with the least expenditure of human energy”. The cost-oriented consumer is defined as one who “bases selection on maximizing the use of money”. It has been demonstrated that the demographics, lifestyle, and consumer values of convenience-oriented consumers are significantly different from those of cost-oriented consumers (Morganosky, 1986).

The increase in productive capacity through technological innovation has resulted in a shift of consumer attitudes from valuing products and services to valuing time (Anderson, 1971). Compared with physical stores, virtual stores have an advantage in term of the ease with which goods can be obtained. Korgaonkar (1984) found that the non-store retailing methods will mainly appeal to convenience and price oriented consumers rather than brand oriented consumers. Especially in recent years, the time needed to obtain an item of merchandise is increasingly a factor in the consumer’s judgment (Korgaonkar, 1984). Convenience has been reported as the primary reason for shoppers to shop on the Internet (Forsythe and Shi, 2003). In addition, information technology and logistic systems have much improved the efficiency of product searching, order taking, and merchandise delivery. Thus, consumers who care more about convenience should have a more positive attitude toward virtual stores.


2.1.3. Impulse tendency
The general trait of impulsiveness has been extensively studied by clinical and developmental psychologists, education researchers, and criminologists (Kacen and Lee, 2002; Rook and Fisher, 1995). Impulsiveness is characterized by unreflective actions (Eysenck et al., 1985), and is significantly correlated with thrill-seeking (Weun et al., 1998), as well as the psychological need to maintain a relatively high level of stimulation (Gerbing et al., 1987; Kacen and Lee, 2002). Puri (1996) proposed that the accessibility of the costs versus the benefits of impulsiveness influences whether consumers behave in an impulsive or a controlled manner. Whether or not an individual focuses on the costs or the benefits of impulsiveness may also depend on an individual’s basic values (Puri, 1996).
Buying impulsiveness is a consumer’s tendency to buy spontaneously, unreflectively, and immediately. Highly impulsive buyers are more likely to experience spontaneous buying stimuli; their shopping lists are more “open” and receptive to sudden, unexpected buying ideas. Furthermore, their thinking is likely to be relatively unreflective, prompted by the physical proximity to a desired product, dominated by an emotional attraction to it, and absorbed by the promise of immediate gratification (Rook and Fisher, 1995). Technologies such as television shopping channels and the Internet expand consumers’ impulse purchasing opportunities, increasing both the accessibility to products and services and the ease with which impulse purchases can be made. Impulse buying generates over $4 billion in annual sales volume in the United States (Kacen and Lee, 2002). The impulse buying tendency is defined as the “degree to which an individual is likely to make unintended, immediate, and unreflective purchases (i.e., impulse purchases)” (Weun et al., 1997). Theoretically, the purchasing process includes five stages, said need recognition, information search, alternative evaluation, purchase decision and post-purchase behavior (Kotler and Armstrong, 1997). Thus, impulsive consumers tend to go through these purchasing stages more quickly. Purchasing environments that help consumers go through purchasing stages quickly will be more attractive to impulsive consumers.
Some characteristics in virtual stores enable consumers to go through the five purchasing stages very quickly. For instance, most virtual stores provide 24-h service all year round, consumers can search for information and purchase products at any time. The virtual stores are reachable for distant customers who can “visit” virtual stores at home and save on transportation time. In addition, virtual stores provide a sufficient purchasing process that combines with direct marketing schemes to attract consumers’ immediate responses. Thus, impulsive consumers may show a more positive attitude toward virtual stores than rational consumers.
H3. Consumer impulse tendency positively affects consumer attitude toward virtual stores.
2.2. The relationship between consumer attitude and behavior
According to Fishbein’s extended model, known as the theory of reasoned action, behavior is determined by intentions, which are in turn determined by attitudes and subjective norms (Ajzen and Fishbein, 1980). The model represents an attempt to combine both individual level and interpersonal/group factors within a single paradigm. However, purchasing goods in virtual stores seems to be mostly determined by individual level factors, as claimed by Netemeyer (1992): “getting a good deal on a product would be viewed as a valued personal consequence.” Many study results suggest that the effect of attitude is stronger than subjective norms (Burnkrant and Page, 1982; Miniard and Cohen, 1979; Trafimow and Fishbein, 1994; Warshaw, 1980). The relationship between attitude and behavioral intentions has been widely examined and supported. Meta-analysis, combining samples of over 10,000 participants, supports the strong attitude–intention–behavior linkage (Kim and Hunter, 1993). Most behavioral models trace causal links from attitude, through intention, to actual behavior, implying that behavioral intentions must be understood to predict behavior from attitudes (Kim and Hunter, 1993). In addition, Armstrong et al. (2000) stated that purchase intentions could provide better forecasts than a simple extrapolation from past sales trends. Berger et al. (1994) recommended that managers’ efforts should be focused on altering consumer attitudes prior to guiding their behavioral decisions.
In current marketing environment, consumers can easily receive lot of information. Retailer managers generally use many schemes to attract and communicate with target consumers to increase sales. If consumers have positive attitudes toward virtual stores, they are more likely to be attracted by the marketing scheme of virtual stores, hence visit the virtual stores and make purchasing decisions. In other words, consumer attitude toward virtual stores is expected to be positively related to consumer purchase intention in virtual stores. Herein, we extend the application of reasoned action theory to virtual stores, and empirically investigate the relationship between consumer attitude and purchase intention toward virtual stores.
H4. Consumer attitude and purchase intention toward virtual stores are positively related.
2.2.1. Conceptual model
In addition to previous factors, consumer’s past behavior is an important antecedent of consumer attitude, as well as future purchase intention. For consumers who have experience of purchasing goods in virtual stores, they should have more positive attitude toward virtual stores than the others. Their likelihood to purchase goods in virtual stores in the near future should be higher than the others. Therefore, consumer’s previous purchase is included in our conceptual model as a control variable. Fig. 1 shows the conceptual model and the hypotheses.

Fig. 1. Conceptual ModelFigure options
3. Methodology
A closed-end questionnaire was designed to collect field data. Four parts were included in the questionnaire. First, to assure that all respondents had a common understanding of virtual stores, the following explanation of virtual stores was provided: “A virtual store denotes a private retailer, without a fixed showroom and face-to-face contact, using information techniques and media to communicate with consumers and pursue marketing goals, such as catalog shopping, TV shopping channel and the Internet store”. After reading this definition of virtual stores, respondents were asked to give their likelihood of purchasing goods in virtual stores. The second part of the questionnaire contained statements intended to measure consumer attitude toward virtual stores, and the third part contained statements to measure risk averseness, convenience orientation, and impulse tendency. These latent constructs were measured by multi-item scales. Respondents were asked to check their degree of agreement with each statement, ranging from strongly agree (7 points) to strongly disagree (1 point). Finally, the respondents were asked for demographic data, and whether they had ever bought goods in virtual stores. Respondents who had previously bought goods in a virtual store were further asked about their satisfaction level with virtual stores, from very satisfied (5 points) to very unsatisfied (1 point).
3.1. The measurement scales
Numerous measurement scales have been drawn and modified in the literature. The risk averseness measures used here were taken and modified from Burton et al. (1998); the convenience orientation measures from Korgaonkar (1984); and the impulse tendency from Weun et al. (1997). The measures of consumer attitude toward virtual stores were taken from Lee (2005), which was established by a procedure proposed by Churchill (1979). After creating an initial item pool, coefficient alpha, item-to-total correlation, and confirmatory factor analysis are used to screen and purify the scales of consumer attitude toward virtual stores (Fornell and Larcker, 1981; Gerbing and Anderson, 1988; Tian et al., 2001). Consumer purchasing intention in a virtual store was measured by asking respondent’s likelihood to purchase in a virtual store, from very likely (7 point) to very unlikely (1 point), which was a single-item scale. All of the above scales were measured by 7-point Likert-type scale. Previous purchase is measured by a dummy variable, “1” represents that consumer has experience of purchasing goods in virtual stores, whereas “0” represents he/she has no experience. Appendix A lists all of the measurement statements of latent constructs.
3.2. The sample
Convenience sampling, instead of random sampling, is applied to collect field data. Theoretically, random selection offers the best chance of minimizing selection effects because each individual in the population has an equal chance of being sampled. In practice, a sample drawn at random is not necessarily representative of the population because not all persons who are approached will agree to participate and refusal rates may vary across different types of individuals. Besides, not all persons who volunteer will complete their participation in the study, and the distribution of those who drop out or fail to complete some measures is also not likely to be random. These problems indicate that even samples drawn at random from a population may be biased in various ways (Hultsch et al., 2002). In addition, the potential consumers of virtual stores include a wide distribution of population, which increases the difficulty of applying random sampling. The purpose of our study is to investigate the relationship between constructs in consumer’s mind, not to estimate the characteristics of the population. A previous study shows that the pattern of relationships is markedly similar across random and convenience samples (Hultsch et al., 2002). As a result, we decide to use convenience sampling which is by far the most common sampling strategy in psychological research (Hultsch et al., 2002). A convenient sample containing 400 working adults in northern Taiwan was invited to participate in the main study. After introducing the purpose of the research face to face, each subject was given a questionnaire and was asked to fill it in. As a result, 377 questionnaires were completely finished and returned, with a 94.3% return rate. The profile of these 377 subjects is listed in Table 1. Among the respondents, 54.1% are female, average age is 32.6 years old, average monthly income is NTD 40,365 (about USD 1280), and 74% of the respondents have purchasing experience in virtual stores.
Table 1. Profile of the sample
Demographic factor
Valid no.
Percentage (%)
Gender Male: 173 45.9

Female: 204 54.1
Education Senior high school and below: 54 14.4

College: 281 74.5

Masters: 42 11.1
Have experience of purchasing in virtual stores or not? Yes: 278 73.7
No: 99 26.3
Satisfaction with virtual storea 1. (Very unsatisfied): 4 1.4

2. (Unsatisfied): 26 9.4

3. (Fair): 97 34.9

4. (Satisfied): 132 47.5

5. (Very satisfied): 19 6.8
Age Mean=32.6

Std. Deviation=9.72
Income (NTD/month) Mean=40,365

Std. Deviation=40,896
Remark: Total sample size N=377.
a
Number of respondents who are experienced in virtual stores: 278.
Table options
4. Result
4.1. Reliability and validity
Table 2 lists the coefficient α, reliability and average variance extracted (AVE) of each factor included in the conceptual model (Fornell and Larcker, 1981). Purchase intention factor is not listed because it is single-item measure. The coefficient α shows good internal consistency of measures, ranging from 0.75 to 0.90 (Nunnally and Bernstein, 1994). The reliability is acceptable across all of the components, ranging from 0.75 to 0.91 (Fornell and Larcker, 1981).
Table 2. Reliability and AVE of latent factors
Factor
Item no.
Coefficient α
Reliabilitya
Average variance extracted (AVE)a
Consumer attitude toward virtual stores 7 0.90 0.91 0.59
Risk averseness 3 0.76 0.80 0.58
Convenience orientation 3 0.75 0.75 0.50
Impulse tendency 3 0.77 0.78 0.54
a
Fornell and Larcker (1981).
4.2. Hypotheses test
To test the postulated hypotheses in Fig. 1, structural equation modeling (SEM) with the LISREL VIII program was applied to estimate the structural coefficients and model fitness. Table 4 shows the results of estimation. The indexes of model fitness reveal that the collected data fits the conceptual model very well (GFI=0.93, AGFI=0.90). Consumer risk averseness is negatively related to consumer attitude toward virtual stores, significant at the α=0.1 level (t=−1.81), so H1 is supported. Consumer convenience orientation is positively related to consumer attitude toward virtual stores, significant at the α=0.05 level (t=6.94), so H2 is also supported. Consumer impulse tendency is positively related to consumer attitude toward virtual stores, significant at the α=0.05 level (t=3.36), supporting H3. Consumer attitude toward virtual stores is strongly positively related to consumer’s purchase intention (t=11.54), which supports H4. Finally, consumer’s previous purchase has positive effect on consumer attitude (t=4.85) and purchase intention (t=2.14), which are significant at α=0.05 level. The result reveals that for consumers who have experience of purchasing goods in virtual stores, they have more positive attitude toward virtual stores than the others, and they are more likely to purchase goods in virtual stores again. One may wonder if there is a direct connection between consumer impulse tendency and purchase intention. A competitive model in which the impulse tendency has a direct link to purchase intention is formed to investigate this question. Study result shows that the direct link between impulse tendency and purchase intention is not significant (γ=0.03; t=0.69), supporting the fact that the indirect effect of impulse tendency on purchase intention, through consumer attitude, is more appropriate.
5. Managerial implications
The findings of this study provide valuable insights for managers of virtual stores. First, consumer risk averseness negatively affected consumer attitude toward virtual stores. When purchasing in a virtual store, consumers must carry higher risks of performance, finance, and safety; and these consumer perceived risks may become a barrier for market extension of a virtual store. Various risk relievers such as a money-back guarantee, well-known brand, and reduced price have been proposed and tested (Dirk and Leunis, 1999), each policy differs in term of cost and efficiency. Managers of virtual stores must seriously evaluate each policy and design their marketing programs to reduce consumers’ perceived risk.
Secondly, consumer convenience orientation positively affects consumer attitude toward virtual stores. In this busy and speedy economy, time saving and convenience have become important advantages of virtual stores. Information technology helps consumers search for products more efficiently, popular home-delivery systems reduce the waiting time and shipping cost, and credit cards make payment more flexible. These factors improve the level of consumer convenience, and hence support the growth of virtual stores. Managers of virtual stores should utilize these factors to improve operational processes and enhance consumer attitude toward virtual stores.
Thirdly, the consumer impulse tendency is positively related with consumer attitude toward virtual stores. New technology provides managers with multiple tools to contact and attract consumers’ attention. At the same time, each consumer receives and deals with lot of information every day, so they are not sufficiently patient to read poorly presented information. As a result, being innovative to attract consumer attention and elicit a consumer response has become an important challenge for managers, which is an extremely important influence on the success of virtual stores.
Finally, consumer attitude toward virtual stores is strongly positively related with purchase intention, which supports the important role of attitude and complies with the theory of reasoned action. The attitude measure used in this study is a valid and efficient scale. Managers of virtual stores could combine with other cross-section market surveys to search for potential customers (consumers who have a more positive attitude toward virtual stores than others). In addition, managers of virtual stores could use this attitude measure to evaluate specific market segmentation and adjust their marketing strategy.
6. Conclusion
Virtual stores have been increasing their importance in marketing channels, but solid knowledge about consumer attitude toward virtual stores remains scarce. Based on reasoned action theory, this article takes the consumer’s point of view to explore the antecedents of consumer attitude by focus group discussion and propose a conceptual model to investigate consumer attitude toward virtual stores and empirically tests the proposed model and hypotheses. The proposed conceptual model is just a beginning, which could be used as a basic model and extended in further research. By improving our understanding of consumer attitude, academics and practitioners can enhance their knowledge and skill for managing various types of virtual stores. However, together with the fast growth of information technology, new types of virtual stores with new business models are continuously being developed and introduced by practitioners; but precisely how consumers react to these changes is still unknown.

Effects of campus foodservice attributes on perceived value, satisfaction, and consumer attitude: A gender-difference approach
Pengarang       : David Joon-Wuk Kwun

Abstract
This study explores a conceptual framework incorporating interrelationships among campus foodservice attributes, perceived value, satisfaction, and effects of these on consumer attitude. Mediating roles of perceived value and satisfaction as well as gender differences in attitude formation process also are investigated within the conceptual framework. Results confirm that enhanced performance in service and product quality, menu, and facility have favorable effects on perceived value, satisfaction and, ultimately, on consumer attitude. However, the attitude-formation process and mediating roles of perceived value and satisfaction differ notably between male and female consumers.

Research highlights
▶ This study explores a conceptual framework incorporating interrelationships among campus foodservice attributes, perceived value, satisfaction, and effects of these on consumer attitude. ▶ Mediating roles of perceived value and satisfaction, and gender differences in attitude formation process also are investigated within the conceptual framework. ▶ Results confirm that enhanced performance in service and product quality, menu, and facility have favorable effects on perceived value, satisfaction and, ultimately, on consumer attitude. ▶ However, the attitude-formation process and mediating roles of perceived value and satisfaction differ notably between male and female consumers.
1. Introduction
As student enrollment increases continuously with diverse cultural and social individualities, more colleges and universities pay increased attention to the importance of their campus foodservices. Overall, college enrollment increased from 14.3 million to 17.5 million between 1995 and 2005 from both under age 25 (33%) and 25 and over (18%) age categories (National Center for Education Statistics [NCES], 2008). In addition to this increase, some 434,000 students attended non-degree-granting postsecondary institutions in fall 2005. From 2005 to 2016, NCES projects a rise of 15% in under-25 enrollments and of 21% in enrollments of 25s and older (NCES, 2008). Within this same period, a significant increase of the female proportion was noticeable, which led an interest to understand the gender differences in academic achievement and quality of campus life. Gender difference, in general, has been identified as one of the important personal characteristics in understanding consumer behavior (e.g., [Liberman et al., 2001], [Meyers-Levy and Maheswaran, 1991] and [McCleary et al., 1994]), including gender similarities and differences in foodservice management and food-choice related studies (e.g., [Blanck et al., 2008], [Driskell et al., 2006] and [Han and Ryu, 2006]).
Additional factors are that college students eat more meals away from home than at home and have various dining experiences from a range of restaurant segments, which raised their expectations of campus foodservices. Food is a critical contributor of physical well-being and a major source of pleasure, worry and stress (Rozin et al., 1999). Gramling et al. (2005) insist that food imposes unique cultural and social functions and relates to general patterns of a specific cultural group, especially in the college-student cohort. Growth in student enrollments and their diverse personal characteristics imposes on colleges and universities the necessity to transform, reposition, and/or introduce new products and services to meet this consumer base’s needs; thus campus foodservice has become a strategic part of this major trend. Campus foodservices are expected to be more responsive to their consumers’ higher expectations and have pushed many institutions to overhaul their campus foodservice operations. More than ever, in the higher education milieu, foodservice became an essential component affecting the quality of campus life and providing strategic and competitive advantages ( [Gramling et al., 2005], [Horwitz, 2005] and [June, 2006]). Consequently, once discreetly considered as an institution’s supporting product – commonly referred to as a meal plan – many colleges and universities try to revamp their campus foodservices to meet diverse and refined consumer needs, by providing greater variety and healthier menus, better service, and socially- and culturally-conscious options.
Nevertheless, perceptions of campus foodservices tend to be unfavorable among this cohort due to various situational, contextual, and environmental constraints. These constraints, such as captive environment, repetitive consumption of limited and monotonous menu items, mediocre execution of food and service, and facility in general ( [Gramling et al., 2005] and [Klassen et al., 2005]), distinguish campus foodservice from other commercial restaurants. Although several researchers have investigated this growing foodservice segment (e.g., [Andaleeb and Caskey, 2007], [Kim et al., 2004] and [Kim et al., 2009]), their approaches generally have focused on the relationships between a set of campus foodservice attributes and satisfaction or behavioral intentions, providing limited understanding of consumers’ emotional evaluations of campus foodservice. Prior studies suggest that consumer attitude is a strong criterion construct in understanding consumers’ summative evaluations of a product or brand and of their behavioral intentions. This stream of research has conceptualized consumer attitude as a relatively permanent and stable evaluative summary of a product or brand captured in such positive or negative dimensions as attractive–unattractive, likable–dislikable, pleasant–unpleasant ( [Ajzen, 2001], [Bolton and Drew, 1991], [Ekinci et al., 2008], [Kraus, 1995], [Gresham et al., 1984] and [Suh and Yi, 2006]). The central role of consumer attitude in prior research has an important ramification in campus foodservice management. Given the varied consumer base, unique constraints, and increased competition from off-campus food providers, understanding consumers’ emotional affects toward a campus foodservice may provide additional insights.
Therefore, building on previous research, this study incorporates consumer-attitude theory to comprehend consumers’ emotive evaluations (i.e., emotional affects) toward a campus foodservice experience. In addition, while the importance of quality, perceived value, and satisfaction is widely recognized in previous studies, further examination of relative importance is needed by investigating these constructs concurrently in an integrated model. Hence, the purpose of this study is to explore a gender difference in attitude-formation process by incorporating the interrelationships among foodservice attributes, perceived value, and satisfaction, and investigates their effects on consumers’ attitudes toward campus foodservice. Within the conceptual framework, perceived value and satisfaction are modeled to mediate the effects of campus foodservice attributes on consumer attitudes. In addition, gender differences in attitude formation process are examined by analyzing male and female consumers separately.
2. Conceptual background
The research framework guiding this study is presented in Fig. 1. The research framework is extended from prior studies on perceived value and satisfaction and their relationships with quality attributes. In addition, the framework expands the scope of previous research on campus foodservice by incorporating the theories pertinent to consumer attitudes and gender differences. This study focuses on major attributes of campus foodservice and their effects on perceived value and satisfaction and, consecutively, on consumer attitude. As shown in Fig. 1, mediating roles of perceived value and satisfaction between campus foodservice attributes and consumer attitude, as well as gender differences in attitude formation process are proposed within the same framework. In summary, this study proposes that consumer-attitude formation processes towards campus foodservice may vary between genders and is a function of campus foodservice attributes (i.e., service quality, food quality, menu, and facility), perceived value, and satisfaction.
2.1. Campus foodservice attributes
2.1.1. Food quality and service quality
Perceived quality is a consumer’s perception, relative to alternatives, of the overall quality or superiority of a product or service with respect to its intended purpose (Zeithaml, 1988). Consumers do not necessarily evaluate quality objectively; instead, evaluation is based on individual perception about what is important to each consumer. In view of that, perceived quality is defined as the consumer’s judgment about an entity’s overall excellence or superiority (Zeithaml, 1988). As a core product of a foodservice operation, food quality has received the most consideration and has been investigated in various aspects, such as flavor, aroma, texture and temperature (e.g., [Andaleeb and Caskey, 2007], [Kim et al., 2009], [Kwun and Oh, 2006] and [Namkung and Jang, 2007]). Increasing interest in and attention to service-quality literature also brought considerable attention to foodservice research and has identified service quality as another core element in dining experience (e.g., [Andaleeb and Conway, 2006] and [Brady et al., 2001]). Both of these tangible and intangible aspects of foodservice operation have been identified as essential attributes in understanding perceived value, satisfaction, and behavioral intentions of consumers ( [Kivela et al., 1999], [Kwun and Oh, 2006], [Namkung and Jang, 2007] and [Pettjohn et al., 1997]), as well as in campus foodservices ( [Andaleeb and Caskey, 2007], [Kim et al., 2004] and [Kim et al., 2009]). Thus, the first four research hypotheses are proposed:
H1. Service quality has a positive effect on perceived value.
H2. Food quality has a positive effect on perceived value.
H3. Service quality has a positive effect on satisfaction.
H4. Food quality has a positive effect on satisfaction.
2.1.2. Menu
This study detaches menu effects from food quality for more in-depth investigations. Whereas, food quality is commonly measured via technical qualities of food, a separate evaluation of menu may reveal additional aspects of campus foodservice. That is, food-quality evaluation most likely is constrained to food which the consumer ordered from available menus. On the other hand, evaluating menu inclines to understanding of consumers’ wants and choices that may or may not be available in a specific restaurant and that expose general trends in foodservice environments. For example, major full service- and fast food-restaurants are changing their menus and providing wider variety, so they can stay abreast of general industry trends and consumer needs (e.g., obesity, health consciousness, lifestyle, demographic changes, etc.) to gain competitive advantage against major competitors (e.g., [Glanz et al., 2007] and [Gregory et al., 2006]). Most noticeable menu trends in campus foodservice include healthy menus, international- and ethnic-flavors, specialty coffees and other beverages, plus promotional- and convenient menus like grab-and-go items (e.g., [Duecy, 2005], [Horwitz, 2005] and [June, 2006]). Kim et al. (2004) and Klassen et al. (2005) confirmed that menus have significant effects on enhancing satisfaction among campus foodservice clientele. Thus, as part of a core element in foodservice management, positive perception on menu (e.g., variety, healthy menu, convenient menu, ethnic menu, and promotional menu) could enhance consumer’s perceived value and satisfaction. This leads to the following hypotheses:
H5. Menu has a positive effect on perceived value.
H6. Menu has a positive effect on satisfaction.
2.1.3. Facility
Facility has been identified as an additional factor affecting customers’ perceptions of a restaurant and campus foodservice (e.g., [Andaleeb and Caskey, 2007], [Auty, 1992], [Kim et al., 2009] and [Kwun and Oh, 2006]). Facility, in this study, refers to the atmosphere and operational characteristics of the sample campus foodservice. Consumers’ expectations of food and dining experience tend to differ, depending on where they consume. Campus foodservices have operational characteristics and atmosphere dissimilar from and more constrained than commercial restaurants. Cardello et al. (1996) asked respondents to rate their expected liking of foods in different locations and found home- and traditional full service restaurants ranked higher than institutional foodservice. While airline- and hospital-food service ranked lower than school foodservice, the latter ranked even below fast food restaurants. When Meiselman et al. (2000) served identical meals in different locations within a university campus, the cafeteria scored lowest compared to the food science laboratory and the dining hall in a training restaurant. Previous studies show several key attributes of facility in campus foodservice had significant effects on satisfactions and revisit intentions, such as cleanliness, dining-room environment, comfort level, operating hours and days, atmosphere, and capacity ( [Andaleeb and Caskey, 2007], [Kim et al., 2004], [Kim et al., 2009] and [Klassen et al., 2005]). Based on related studies and similar reasoning for preceding campus foodservice attributes, favorable evaluation on campus foodservice facility is expected to have positive effect on perceived value and satisfaction. Hence,
H7. Facility has a positive effect on perceived value.
H8. Facility has a positive effect on satisfaction.
2.1.4. Perceived value and satisfaction
Perceived value and satisfaction have received much attention in service- and marketing-literature and extensively have been recognized as one of the salient constructs in understanding consumer behavior. The significant roles of these constructs in consumer research have provided much theoretical debate and empirical data, but still signify further need to investigate their interrelationships collectively. Several seminal studies (e.g., [Cronin et al., 2000] and [Oh, 1999]) provided foundation for this stream of holistic approach that integrates quality, perceived value, and satisfaction within a same research framework. Based on the means-end approach proposed by Zeithaml (1988), previous studies associated consumers’ perceptions on product and/or service quality attributes to perceived value and satisfaction, and emphasized the conceptual interrelationships and their significant effects on loyalty and behavioral intentions (e.g., [Brady et al., 2001], [Cronin et al., 2000], [Gallarza and Saura, 2006], [Hutchinson et al., 2009] and [Oh, 1999]). Extended from previous literature, this study integrates foodservice attributes, perceived value, and investigates their effects on consumer attitude toward campus foodservice.
Following Zeithaml’s (1988) widely accepted conceptualization, perceived value is defined as the customer’s overall assessment of the utility of a product based on perceptions of what is received and what is given. This study views satisfaction as a transaction- and performance-based evaluation process resulting from discrepancies between prior expectation and perceived performance (e.g., [Dubé and Morgan, 1998] and [Westbrook and Oliver, 1981]). Previous studies hold that improvement in quality, value and satisfaction in a service encounter enhances favorable outcomes. In general, those studies provided common consensus of contributory sequence from quality, value and then satisfaction, which subsequently affects consumers’ behavioral intentions (e.g., [Brady et al., 2001], [Cronin et al., 2000], [Gallarza and Saura, 2006], [Hutchinson et al., 2009] and [Oh, 1999]). In this study context, consumers’ subjective and relative evaluations on campus foodservice attributes provide the basis for the give-and-take components of perceived value. This implies that perceived value is a trade-off received from all relevant benefits from campus foodservice attributes and it sequentially precedes satisfaction. Based on the above argument and findings, this study proposes the following hypothesis:
H9. Perceived value has a positive effect on satisfaction.
2.2. Consumer attitude
Attitude, generally defined as summative evaluations of a product or brand, has provided a wealth of cumulative contributions in marketing and consumer research. Fishbein and Ajzen (1975) posit that individuals respond to an object (or an idea) or a number of things (or opinions) and explore the construct of attitude as a learned predisposition of humans. General agreement exists that attitude is an evaluative judgment about objects, which represents a person’s enduring favorable or unfavorable evaluations and emotional feelings guiding action tendencies toward those objects. Considerable research suggests that consumer attitude is a strong criterion construct in understanding consumers’ summative evaluations of a product or brand and of their behavioral intentions (e.g., [Ajzen, 2001], [Bolton and Drew, 1991], [Kraus, 1995], [Gresham et al., 1984] and [Fishbein and Ajzen, 1975]). As mentioned earlier, campus foodservice generally has a mixture of situational, contextual, and environmental constraints (e.g., captive environment, limited and repetitive menu, and dining time limitation). In addition, college students have demanding lifestyles (e.g., work, study, and family) and are socially-sophisticated as well as culturally-diverse consumers (e.g., ethnic background, nationality, and age category). In view of that, campus foodservice has become, to a greater extent, a place not only to eat, rest, and study, but also to enhance the sense of community and quality of campus life. Hence, understanding consumer’s emotional affect with continuous attitude may provide additional insights toward campus foodservice experiences.
Oliver (1980) distinguishes between antecedent attitude and continuous attitude based on the prepurchase and postpurchase service evaluations, and posits that satisfaction influences future behavioral intention as well as postpurchase attitude. Based on this conceptualization, this study incorporates the continuous attitude and views it as a summative and emotive evaluation toward campus foodservice experiences at the post-consumption stage. In line with Oliver’s (1980) study, Ekinci et al. (2008) focus on the continuous attitude and insist that satisfaction is a better indicator of the consumers’ overall attitude than service quality to hospitality firms. Accordingly, satisfaction is different from consumer attitude, which mediates the relationship between antecedents (i.e., actual and ideal self-congruence, desire congruence, and service quality) and intention to return and consumer’s continuous attitude. Previous studies differentiate satisfaction from consumer attitude, and recognize satisfaction as an essential antecedent of overall attitude to a product or brand at the post-purchasing point (e.g., [Bolton and Drew, 1991], [Ekinci et al., 2008], [Oliver, 1980] and [Suh and Yi, 2006]). Based on these theoretical foundations and empirical studies, this study proposes the following hypotheses:
H10. Perceived value has a positive effect on consumer attitude.
H11. Satisfaction has a positive effect on consumer attitude.
2.3. Gender differences
The significant role of females in our society corresponds to their large purchasing power and increase in female business-travelers for the past several decades. In addition, a noticeable trend in college enrollment is that the gender gap in college participation has reversed from 2003; much of the student-enrollment growth between 1995 and 2005 mentioned earlier is female. Accordingly, the number of females enrolled rose 27%, the number of males only 18%. In addition, 51% of women had entered and/or completed college compared to 41% of men (NCES, 2005). Considering this disproportion, the importance of food in general and its consumption trends, gender differences may play an essential role in campus foodservice management perceptions and effects.
This trend led a growing interest among hospitality and tourism researchers to delineate gender similarities and differences in consumption behaviors. McCleary et al. (1994) studied the differences between male and female business travelers in hotel selection and found that security, personal services, and low price were more important to female- than to male-travelers. Mattila (2000) investigated gender differences and consumer evaluations of service encounters but found no significant differences between genders. According to Oh et al. (2002), male and female travelers differed significantly on expectations and perceptions of lodging services, while there were no notable differences on satisfaction and behavioral intentions. In Han and Ryu’s (2006) study, gender differences showed a significant moderating role in the relationship between customer satisfaction and revisit intention in an upscale restaurant; female customers showed a stronger intention to revisit the restaurant when satisfied than did male customers.
According to selectivity theory, males often do not engage in comprehensive processing of all available information as a basis for judgment but, instead, use selective cues which are highly available and salient in the focal context. On the other hand, females attempt to engage in effortful, comprehensive and itemized analysis of all available and accessible cues ( [Meyers-Levy, 1989] and [Meyers-Levy and Maheswaran, 1991]). In addition, females often are more attuned to their emotional states and assign more value to such feelings to arrive at buying decisions than do males (e.g., [Dubé and Morgan, 1998] and [Wood, 1989]). Correspondingly, personal interaction processes with service providers strongly influence female purchasing behaviors ( [Homburg and Giering, 2001] and [Gilber and Warren, 1995]). Parallel to these views on gender differences, Holbrook (1986) insists that females show more tendencies to be visually oriented, intrinsically motivated, and romantic than do males, and suggests that gender differences may be a key variable in moderating consumers’ evaluative judgments.Gender differences have been investigated widely in food-choice related studies. In general, studies have shown that females pay more attention to, and have more concern about, food choice than do their male counterparts (e.g., [Blanck et al., 2008], [Liberman et al., 2001] and [Unklesbay et al., 1998]). Rozin et al. (1999) found a noticeable difference between genders in attitude to food and to the role of food in life. Accordingly, females tend to worry about food and are health oriented, compared to males who value pleasure and have culinary-oriented attitude toward foods. Driskell et al. (2006) studied the eating habits of university men and women at fast-food restaurants. They found that a higher percentage of men eat fast foods more often and ordered larger portion sizes with carbonated beverages than did women. Nevertheless, most gender-difference studies in food choice research were mainly descriptive, providing limited understanding of its compelling effects on consumption behavior. Previous studies insist that consumers provide summary judgments of their dining experiences by relying on a set of selected attributes and ascribing different levels of importance to each attribute (e.g., [Auty, 1992], [Kim et al., 2009], [Kivela et al., 1999] and [Namkung and Jang, 2007]). These discussions suggest that males and females may evaluate campus foodservice attributes in differing ways and, therefore, may show different results on perceived value, satisfaction and, eventually, on consumer attitude. Selectivity theory and results from previous studies on gender difference imply that female consumers may consider a much broad range of campus foodservice attributes in attitude formation process than male counterparts. This reasoning leads to the following hypothesis:
H12. Consumer’s attitude formation process is different between genders.
3. The study
3.1. Questionnaire development
A self-administered questionnaire comprising three major parts was composed with measurement items that were used in the literature or slightly modified for the purpose of this study (see Table 2). Part One was designed to understand consumers’ evaluations on service quality (6 items), food quality (6 items), menu (5 items), and facility (7 items) aspects of the campus foodservice attributes. Reviews of related studies (e.g., [Andaleeb and Conway, 2006], [Kim et al., 2004], [Kim et al., 2009], [Kivela et al., 1999], [Klassen et al., 2005] and [Namkung and Jang, 2007]) provided measurement items for this part. Respondents were asked to evaluate on five-point scales ranging from poor (1) to excellent (5). Part Two was aimed at identifying consumers’ evaluations on perceived value, satisfaction, and consumer attitudes toward the campus foodservice (e.g., [Andaleeb and Conway, 2006], [Gallarza and Saura, 2006], [Kim et al., 2009], [Klassen et al., 2005], [Oh, 1999] and [Westbrook and Oliver, 1981]). Perceived value was measured with four items and satisfaction was measured with a single item with five-point scales ranging from strongly disagree (1) to strongly agree (5). Consumer attitude was measured using four 5-point semantic scales anchored by displeasing/pleasing, not likeable/likeable, not enjoyable/enjoyable, unattractive/attractive. Part Three was designed to measure demographic and other information. Before the questionnaire was finalized, minor revision on measurement items and wordings were made based on reviews from convenience sample of six professors and graduate students who were familiar with the scope of this study.
3.2. Data collection
This study used a web-based survey to collect data. Target sample for this study was approximately 2500 undergraduate and graduate students, faculty, and professional staff employed at one of the largest hospitality colleges. The college is located in a satellite distant from its main campus and has a campus foodservice venue with 150 seats and coffee shop, several grab-and-go sections, and five food-court restaurants. Formerly contract-managed by a third party college-foodservice operator, the college assumed its responsibility and has been providing meal accommodations. Data was collected several weeks before final exams (i.e., from mid March to mid April). First, a brief ‘prenotice email’ and purpose of the study was sent a few days prior to the delivery of the online survey. Second, the questionnaire was distributed with the objectives of the study and confidentiality of the participant’s response. Third, two e-reminders were sent within the data-collection period.
3.3. Data analysis methods
Data analysis for this study was based on three major processes. First, exploratory factor analysis was utilized to reduce the 32 measurement items into a parsimonious data structure. Second, an independent sample t-test was conducted to initially evaluate the mean differences of those factors and other outcome variables between genders. Third, a series of multiple regression models were used to test proposed conceptual relationships. To examine the gender difference in attitude formation process, the respondents were divided into groups of male and female and were analyzed separately. Further consideration was given to mediating effects of perceived value and satisfaction between dimensional perceptions of campus foodservice attributes and consumers’ attitudes toward a campus foodservice. These sequential regression models were based on the conceptual procedure proposed by Baron and Kenny (1986). The sequence used was: (1) regress the mediators on the independent variables, (2) regress the dependent variables on the independent variables, and (3) regress the dependent variables on both the independent variables and mediators.
4. Results
4.1. Sample profile
Within the data collection period, 440 usable responses were gathered from the approximately 2500, yielding an 18% response rate. Overall, demographic information of the respondents closely represented the study population in terms of age category, status, gender and ethnic background. As indicated in Table 1, the sample comprised 107 male respondents (24.3%) and 333 female (75.7%). Among these 440 respondents, 322 (73.2%) were categorized as White/Caucasian, 47 (10.7%) as Hispanic/Latino, 29 (6.6%) as Asian, 19 (4.3%) as African American, 19 (4.3%) as Native American, and 4 (0.9%) as Hawaiian/Pacific Islander. As expected in a typical metropolitan university, a majority of respondents (81.9%) lived off-campus. Approximately 76% of respondents were younger than 25 years old, mostly academic seniors (34.2%) and juniors (34.6%).
Table 1. Demographic information.
Demographics

Age (mean = 23.22)

Ethnicity

Status

Gender <20 td=""> 40 (8.5%) White/Caucasian 322 (73.2%) Freshman 15 (3.4%)
Male 107 (24.3%) 20–24 321 (67.7%) African American 19 (4.3%) Sophomore 52 (11.7%)
Female 333 (75.7%) 25–29 42 (8.9%) Asian 29 (6.6%) Junior 154 (34.6%)
Residence 30–34 13 (2.6%) Hispanic/Latino 47 (10.7%) Senior 152 (34.2%)
On campus 80 (18.1%) 35–40 10 (2.0%) Hawaiian/Pacific Islander 4 (0.9%) Graduate student 52 (11.7%)
Off campus 363 (81.9%) >40 16 (3.3%) Others 19 (4.3%) Faculty/staff 20 (4.5%)
Note: Percentage in parentheses.
4.2. Factor analysis
Exploratory factor analysis with a varimax rotation was utilized to reduce measurement items into a more communicable construct (see Table 2). Portion size was co-loaded in another factor with relatively low factor loading in the first analysis, therefore was deleted for a better factor structure. As a result, the 32 attributes were summarized into six main factors, which captured 77.73% of total variance. This result was consistent with expectation, based on the nature of the measurement items, except for the facility factors. Unlike previous studies ( [Kim et al., 2004], [Kim et al., 2009] and [Kwun and Oh, 2006]) where atmosphere and operation/convenience related measurements were grouped in separate factors, results of the factor analysis in this study combined those into one factor. This result can be appropriate for this study because it is based on a campus foodservice at a hospitality college separately located from its main campus. For that reason, the measurement items were consisted of several operation- and facility-related items rather than with convenient location among multiple campus foodservices within a same campus. While the first factor, service quality, consisted of 6 service-quality-related items with an eigenvalue of 5.38, other factors were the facility (eigenvalue = 4.52), menu (eigenvalue = 3.88), food quality (eigenvalue = 3.79), perceived value (eigenvalue = 3.27), and consumer attitude (eigenvalue = 3.26), respectively. The within-factor Cronbach’s alpha of reliability ranged from .90 to .96, indicating that the derived within-factor measurement items of each constructs were internally consistent. Based on these results, the within-construct items were averaged for use in subsequent analyses to simplify the analyses and interpretation.
4.3. Preliminary data analysis
First, preliminary analysis was performed to understand correlations among all measures and their mean scores and standard deviations. As indicated in Table 3, all constructs were intercorrelated with each other and showed relatively strong correlations with perceived value, satisfaction, and consumer attitude. The factors’ raw mean scores of consumer attitude (mean = 4.37, SD = 1.54), facility (mean = 3.51, SD = 0.84), and service quality (mean = 3.41, SD = 1.02) were relatively higher than those of the food quality (mean = 2.90, SD = 0.96), menu (mean = 2.70, SD = 0.92), perceived value (mean = 2.56, SD = 1.00), and satisfaction (mean = 2.43, SD = 1.09). Although the mean scores of consumer attitude and service quality factors were relatively higher than those of other factors, they also had high standard deviations which indicate that perceptions vary among participants. Both low mean and standard deviation scores of food quality, menu, and perceived value factors were noticeable, signifying some problematic issues of campus dining. Facility was the only factor with relatively high mean score and low standard deviation compared to other factors, while satisfaction showed low mean score with high standard deviation.

4. Discussions
Results of this study provided an additional step toward an integrated research framework of campus foodservice attributes, perceived value, and satisfaction. Specifically, this study investigated the mediating roles of perceived value and satisfaction between foodservice attributes and consumer attitudes toward a campus foodservice. In addition, gender differences were evaluated by analyzing the research framework separately between male and female respondents on a series of sequential regression analyses. Different from previous campus foodservice research that focused on consumer satisfaction, this study proposes several precursors affecting consumer attitudes toward a campus foodservice. In summary, this study’s results provide additional insights to what previous studies have reported, namely that: (1) performance in service and product quality, menu, and facility had different effects on perceived value, satisfaction, and consumer attitude toward campus foodservice; (2) perceived value and satisfaction had different mediating roles between a range of campus foodservice attributes and consumer attitudes; and (3) consumer attitude-formation-process in general was dissimilar between male and female consumers.
First, campus foodservice could be evaluated effectively in four major dimensions which show significant explanatory powers on perceived value and satisfaction, consequently on consumer attitudes. Campus foodservice operations need to appease students, faculty and staff, and the college also must sustain financial stability. Administration and foodservice mangers’ abilities to allocate resources properly based on campus foodservice attributes, perceived value, and satisfaction may enhance overall consumer attitudes toward campus foodservice experience. Administration and foodservice operators need to pay closer attention to enhancing service and product quality, menu, and facility of the campus foodservice to provide a favorable dining experience to socially-sophisticated and culturally-diverse consumers.
Explicit considerations on gender differences are required, as findings suggest that female consumers evaluated their campus foodservice experience differently from how their male counterparts did. Female consumers, in general, evaluated campus foodservice based on variety of attributes, while male satisfaction on campus foodservice derived predominantly from food quality and perceived value. In particular, service quality and menu were also important for female consumers in addition to food quality and perceived value attributes. Correspondingly, food quality and perceived value were most important in the male consumer attitude-formation process; all variables were significant on consumer attitude in the female group, indicating a comprehensive attitude-formation process among female consumers.
Results also showed interesting mediating roles of perceived value and satisfaction between campus foodservice attributes and consumer attitude. Despite the fact that complete mediation effects of perceived value and satisfaction occurred only in some measures, noticeable difference between males and females could be identified. Whereas perceived value completely mediated the effects of service quality on consumer attitude in the female group (Model 5-1), the effects of facility on consumer attitude in the male group were completely mediate by perceived value (Model 5-2). While perceived value was completely mediated by satisfaction in both genders, perceived value and satisfaction had partial mediation effects on other significant foodservice attributes on consumer attitude, respectively. Different roles of service quality and food quality were noticeable also. The effects of food quality need to be emphasized since it had both direct and indirect effects on perceived value, satisfaction, and consumer attitude. On the other hand, service quality was not significant in any regression models among male consumers, while it was one of the most significant attributes in the female group.
Results of gender differences in attitude formation process aligned with the selectivity theory and the importance of emotions and services among female consumers. Accordingly, males tend to focus on few selective cues which are highly available and salient, while food quality and perceived value turned out to be the most important attributes in this study context. On the other hand, attitude formation among females was based on comprehensive processing of all available information in campus foodservice. Noticeably, the facility factor had no effect on perceived value and satisfaction but was exceedingly significant on consumer attitude in both genders in different ways. In particular, male consumers considered facility as one of the trade-off components of perceived value. That is, the effect of facility on consumer attitude was mediated completely by perceived value among male respondents. On the other hand, the facility attribute not only had direct effects on consumer attitude in the female group, but also was the most prominent variable with the strongest standardized regression coefficient on consumer attitude (Model 6-1). Furthermore, the significant effects of service quality on attitude formation process were present only in the female group, corroborating the importance of personal interaction in female consumption processes.
5. Limitations and suggestions for future research
Results of this study provide a foundation for several directions in future research. First, the role of consumer attitude needs further exploration, especially regarding the effect on campus foodservice and other considered strategies in higher education. Although consumer attitude is recognized as one of the influential antecedents of future behavior, several studies signify that additional explanations (e.g., price, quality, convenience, and brand familiarity) need to be included to be a powerful predictor of behavioral intention or marketplace behavior ( [Ajzen, 2001] and [Kraus, 1995]). Especially for college students, campus foodservice is not merely where to consume meals, but has become an essential component of their quality of life on campus. Thus, the importance of food consumption and consumer attitude toward the foodservice facility in general, need to be further realized in understanding their impact on behavioral intentions, overall college experience, and other competitive advantages (e.g., student recruitment and enrollment, institutional reputation, and sense of belonging).
Second, more elaborative understanding of campus foodservice attributes is needed. The major attribute categories of service quality, product quality, menu, and facility showed different effects on perceived value, satisfaction, and eventually on consumer attitude toward campus foodservice. In addition, perceived value and satisfaction showed different mediating effects between foodservice attributes and consumer attitude. These results imply that male and female consumers’ evaluations on campus foodservice attributes provide mixed trade-off benefits on perceived value and different effects on satisfaction and consumer attitude. In particular, different roles of food quality and service quality in general, as well as noteworthy effects of facility on consumer attitude entail further research. In addition, socio-cultural changes and consumers’ refined lifestyles demand more studies in and of menu trends and menu varieties, in addition to operational-related aspects of campus foodservice.
Third, inclusions of additional variables as well as replications of other types of institution and hospitality and tourism sectors are needed. Although the sample product is consist of variety of food choices (e.g., coffee shop, grab-and-go sections, and five food-court restaurants) with 150 seats, this study was based on a campus foodservice in a hospitality-specific college located apart from its main campus. In addition, contrary to general trends in campus foodservice operations, the sample product does not have any branded- or multi-unit foodservice chains in campus foodservice. Results may differ in other types of campus foodservice where a range of specialty coffee chains, fast food restaurants and upscale foodservices provide food and beverage to their campus consumers. In addition, better conceptualization of major constructs integrated in this study also is needed. Those conceptually abstracted variables integrated in this study (i.e., foodservice attributes, perceived value, satisfaction and consumer attitude) are correlated highly and need to be further differentiated from each other with clear conceptual definitions and measurements; this is an ongoing discussion among marketing- and consumer researchers.
Finally, gender difference results, in particular, communicate a need for further investigation. The entire consumer attitude-formation process in this study showed interesting dissimilar results between male and female consumers. Gender differences and other personal characteristics need further exploration not only in the campus foodservice setting, but also in various hospitality and tourism sectors. Against a relative deficiency in previous research, this study provided systematic differences on attitude-formation processes of both genders, measured with a biologically-dichotomous categorization. Although this biological gender categorization provides clear distinction on both genders’ consumption behaviors, other approaches also can be applied in future research. In line with current socio-cultural trends, consumers’ perceptions of gender roles and masculine and feminine personalities also may provide additional insights into the degree of cognitive, affective, and behavioral differences in product and brand evaluations.


Consumerattitude and purchase intention toward green energy brands: The roles of psychological benefits and environmental concern
Pengarang       : Patrick Hartmann &Vanessa Apaolaza-Ibáñez
Abstract
This paper suggests that advertising campaigns directed at increasing consumer demand for green energy should emphasize not only environmental concern and utilitarian benefits, but also psychological brand benefits. The theoretical framework proposes three distinct psychological benefit categories potentially enhancing consumerattitudes toward green energy brands and increasing purchase intentions: warm glow, self-expressive benefits, and nature experiences. A sample of 726 consumers was exposed to experimental advertisements for a fictitious green energy brand. Findings confirm most predicted effects and underline the overall significance of psychological brand benefits. Only self-expressive benefits do neither affect participants’ attitudes toward the experimental brand nor their purchase intentions. Nature experience has the strongest influence on brand attitude. Multi-group structural analysis shows that the nature experiences level evoked by the advertisements moderates the effects of the behavioral antecedents studied on brand attitude and purchase intention. The findings provide keys to improving green energy branding and advertising strategy.

1. Introduction
After the 1970s oil crisis, public awareness of energy related issues has attracted the attention of consumer researchers (e.g., McDougall, Claxton, & Ritchie, 1981). Early findings suggest that environmental awareness encourages consumers to decrease their energy consumption (Kasulis, Huettner, & Dikeman, 1981) and to adopt solar energy (Labay & Kinnear, 1981). “Green energy” or “green power” is derived from renewable energy resources, including photovoltaic and thermoelectric solar energy, biomass, geothermal and wind energy. Currently, some consumers pay a premium price for branded green electricity provided by, for example, Green Mountain Energy (U.S.), Ecotricity (U.K.), Lichtblick (Germany), NaturEnergie (Austria), or Iberdrola Energía Verde (Spain).
Higher generation costs and the consequently higher market prices constitute the principal barrier to consumers’ adoption of green energy (Salmela & Varho, 2006). Public opinion surveys find that up to 30% of consumers are willing to pay a price premium for green energy ( [Eurobarometer, 2003], [Eurobarometer, 2005] and [Zarnikau, 2003]). However, to date, green energy brands’ market share remains low (Gan, Eskeland, & Kolshus, 2007) and costs 20% more than regular electricity charges discourage most potential consumers (Salmela & Varho, 2006). Green energy’s future success depends on effective branding and marketing communications strategies designed to enhance consumers’ benefit perception ( [Roe et al., 2001] and [Truffer et al., 2001]). While technical characteristics and green electricity labeling deliver utilitarian benefits to consumers, purchasing green energy potentially derives psychological benefits too. This paper analyzes influences of consumers’ environmental concern and perception of green energy brands’ benefits on attitude toward the brand and purchase intention. The literature review identifies three distinct psychological benefits potentially affecting behavioral intentions: warm glow feelings derived from the moral satisfaction of contributing to the common good environment; self-expressive benefits from conspicuous environmentally sound consumption; and nature experiences evoked by natural brand imagery. The empirical study exposes consumers to experimental advertisements for a fictitious green energy brand measuring utilitarian and psychological benefit perceptions, attitudes toward the brand, and intention to purchase. Structural equation analysis examines the relationships proposed in the theoretical framework.
2. Attitudes toward green energy and the environment
An increasing volume of research addresses cultural, social, and psychological factors in consumers’ demand for green electricity ( [Clark et al., 2003] and [Vringer et al., 2007]). Despite visual impacts of wind turbines (Groothuis, Groothuis, & Whitehead, 2008), attitudes toward green energy are overall favorable globally, contributing to a growth in consumers purchasing premium-priced green electricity ( [Ek, 2005], [Hansla et al., 2008] and [Salmela and Varho, 2006]).
Behavioral effects of a consumer’s personality traits and general environmental attitudes suggest that values and environmental concern are principal determinants of environmentally sound consumption ( [Balderjahn, 1988] and [Diamantopoulos et al., 2003]). Consumers engage in conservation behavior because they are intrinsically concerned about the environment and society ( [Bamberg, 2003] and [Fransson and Gärling, 1999]). Researchers use a variety of alternative and complementary measurement scales to assess consumers’ concern with environmental issues (e.g., [Kinnear et al., 1974] and [Synodinos, 1990]), including the New Environmental Paradigm (NEP) scale ( [Dunlap and Van Liere, 1978] and [Van Liere and Dunlap, 1981]). Several studies confirm that consumer’s environmental concern influences purchase behavior of environmentally sound products ( [Balderjahn, 1988] and [Roberts and Bacon, 1997]). Sensitivity to climate-change issues, awareness of clean energy and alternative energy sources, as well as energy conservation constitute explicit dimensions of environmental concern (Zimmer, Stafford, & Stafford, 1994). Research also shows that green energy consumers are more environmentally concerned than the general population ( [Clark et al., 2003], [Ek, 2005] and [Hansla et al., 2008]). Overall, concern for the natural environment plays a significant role in green energy purchase decisions. Applying the theory of reasoned action (Fishbein & Ajzen, 1975) to the case of green energy, attitudes toward renewable energy mediate the effect of environmental concern on purchase intention (Bang, Ellinger, Hadjimarcou, & Traichal, 2000). Hansla et al. (2008) provide evidence of environmental concern’s direct and indirect effects on consumers’ willingness to purchase green electricity at a premium price.
H1a. Environmental concern influences the intention to purchase green-branded energy positively.
H1b. Attitude toward the brand partially mediates the effect of environmental concern on purchase intention.
Actual energy consumption patterns often diverge from stated concerns about the environment (Vringer et al., 2007). Consumers purchase premium priced green energy only if they perceive sufficient additional benefits (Roe et al., 2001). The following sections discuss how consumers’ perceptions about benefits of green-branded energy may affect brand attitude and behavioral intentions. Both utilitarian and psychological benefits seem to impact purchase decisions.
3. Utilitarian benefits of green energy
Consumers perceive that the consumption of products with environmentally sound attributes (e.g., greener production) delivers additional benefits compared to conventional alternatives ( [Bech-Larsen, 1996] and [Sriram and Forman, 1993]). Many consumers believe that green energy prevents or decelerates climate change and global warming, increases air quality, and decreases energy dependency (Roe et al., 2001). Clark et al. (2003) find that green energy brand adopters perceive green electricity to be more environmentally friendly, lowering future solar energy costs, and reducing reliance on imported oil. Study participants also believe that reducing air pollution from electricity production would improve the health of natural ecosystems and individuals, and that decreasing carbon dioxide emissions would slow global warming. Wüstenhagen and Bilharz (2006) suggest that green power customers intend to contribute to climate protection and renewable energy growth and to ensure that their purchasing decision does not support unsustainable energy sources. Roe et al. (2001) show that consumers who actually paid a price premium for green energy did so particularly with the intention to support the installation of new renewable generation capacity. To enhance perception of utilitarian benefits of green electricity, Salmela and Varho (2006) argue that consumers need a certain amount of information about the environmental impact of different electricity products. Studies confirm that information about environmentally relevant utilitarian product attributes affects purchase intentions ( [Roberts, 1996] and [Scholder-Ellen, 1994]). Exposure to information about energy resource issues increases intention to pay a price premium for renewable energy (Zarnikau, 2003). Green energy labeling helps consumers to identify electricity products with genuine environmental benefits (Truffer et al., 2001); but the current information may be insufficient to inform consumer choice adequately. For example, purchasing decisions depend on whether the potential customer receives information about energy sources only, or also about emission levels (Johnson & Frank, 2006). More accurate and detailed labeling information than typically provided may be necessary to guide consumers’ decision making toward green energy.
The theory of reasoned action (Fishbein & Ajzen, 1975) provides a framework to analyze how perceptions of utilitarian environmental benefits affect green energy purchase intentions. Bang et al. (2000) support the proposition that beliefs about renewable energy relate positively to the intention to pay a price premium for green energy products.
H2a. Perceptions about green energy brand’s utilitarian environmental benefits positively influence purchase intention.
H2b. Attitude toward the brand partially mediates the effect of green energy brand’s utilitarian environmental benefits on purchase intention.
Purchasing green electricity delivers rather limited utilitarian benefits at the individual level. Reducing global warming or energy dependency only becomes a collective benefit when widespread renewable energy adoption occurs. However, green energy brands also offer psychological benefits to consumers.
4. Psychological benefits of green energy brands
4.1. Warm glow
Classical pro-social behavior theory posits that pure altruism motivates individuals to contribute to the common good (e.g., Bergstrom, Blume, & Varian, 1986). The literature conceptualizes altruism as a personal value structure with significant influences on behavior ( [Schwartz and Bilsky, 1987] and [Stern et al., 1995]). However, studies on contingent valuation analysis of the utility of contributing to public goods show that pure altruism does not entirely explain pro-social behavior ( [Andreoni, 1989] and [Andreoni, 1990]). Consumers experience a direct, personal benefit arising from the contribution and independent of any increase in the common good, which Andreoni calls the “warm glow of giving”. With regard to environmentally responsible behavior choices, consumers experience the intrinsic warm glow feeling of well being as a consequence of the moral satisfaction engendered by contributing to the environmental common good ( [Kahneman and Knetsch, 1992], [Nunes and Schokkaert, 2003] and [Ritov and Kahneman, 1997]). This conceptualization is consistent with empirical findings suggesting that some consumers purchase green energy at a premium price in order to feel better with themselves rather than the decision’s environmental impact (Wüstenhagen & Bilharz, 2006). While society as a whole receives benefits from green energy, users experience additional personal warm glow benefits contributing to climate protection and energy independence (Menges, Schroeder, & Traub, 2005). The expectation of warm glow potentially motivates purchase intention either directly or mediated by attitude formation toward the brand.
H3a. Warm glow derived from contributing to the environmental common good positively influences the intention to purchase green-branded energy.
H3b. Attitude toward the brand partially mediates warm glow’s effect on purchase intention.
4.2. Self-expressive benefits
Signaling theory and the literature on symbolic and conspicuous consumption provide a conceptual framework to understand the psychological benefits derived from self-expressive ( [Aaker, 1999] and [Aaker, 2002]), socially visible consumption of environmentally friendly products. Signaling is the process of conveying information about oneself implicitly, by engaging in behaviors that reveal personal traits and preferences to observers. Individuals are more willing to consume in a way that benefits society when signaling is likely (Glazer & Konrad, 1996). Products with a higher signaling potential deliver greater benefits from association with pro-social behaviors (Bennett & Chakravarti, 2009). Product symbolism and symbolic consumption research support this conceptualization ( [Belk et al., 1982] and [Hirschman, 1981]). Most consumer products carry a symbolic meaning often affecting purchase and use ( [Holbrook and Hirschman, 1982] and [Sirgy, 1985]). Solomon (1983) argues that products are relevant for “setting the stage” for consumer’s social roles. Individuals evaluate and place others in a social context to a significant degree by the products they consume. Consumers may conspicuously consume environmentally friendly products in order to display pro-environmental attitudes. However, individuals may also engage in conspicuous environmentally sound behavior to signal their altruism. Conspicuous altruism enhances status and reputation by showing an individual’s capacity and willingness to contribute to the common good ( [Van Vugt et al., 2007] and [Roberts, 1998]). Griskevicius, Tybur, and Van den Bergh (2010) demonstrate that status motives lead consumers to choose green products over non-green alternatives.
Green energy customers should experience psychological benefits from signaling their pro-social and pro-environmental orientation, as well as their capacity to incur extra costs for the sake of the environment and society. Self-expression, as a psychological motive, may induce consumers to purchase green-branded electricity. Brand attitude potentially mediates this effect. Psychological reward expectations from purchasing a green energy brand should enhance consumer’s attitude toward this brand.
H4a. The expectation of self-expressive benefits derived from conspicuous consumption of green-branded energy positively influences purchase intention.
H4b. Attitude toward the brand partially mediates the effect of self-expressive benefits on purchase intention.
4.3. Nature experiences
Most print advertisements and television commercials for green energy brands display visual images of pristine natural scenery. Pictures influence formation of brand beliefs and affective responses to advertisements (e.g., [Mitchell and Olson, 1981] and [Rossiter and Percy, 1980]). Images of natural environments may help construct positive product attribute beliefs, increasing the salience of environmentally sound product features. On the other hand, positive emotional responses evoked by advertising enhance brand affect ( [Burke and Edell, 1989] and [Edell and Burke, 1987]). Natural advertising imagery may exert affective influences as well. Environmental psychology demonstrates that experiencing nature engenders positive emotional responses ( [Hartig et al., 1991], [Kaplan, 1995] and [Ulrich, 1981]). Photographs or video recordings elicit similar affective responses to “genuine” nature experiences ( [Hull and Stewart, 1992], [Kaplan and Kaplan, 1989] and [Nassauer, 1982]). Natural imagery embedded in advertising potentially emulates the effects of nature. Positive emotional responses to advertising result in a more positive brand attitude affecting intention to purchase ( [Batra and Ray, 1986] and [Edell and Burke, 1987]). Positive affect evoked by advertising-induced nature experiences leads to brand attitude improvement (Hartmann & Apaolaza-Ibáñez, 2009) and, indirectly, increases purchase intention. However, influences of affective responses to advertising on intention to purchase the brand may be direct (e.g., [Allen et al., 1992] and [Mitchell and Olson, 1981]), suggesting that hypothesized effects are both direct and mediated by brand attitude.
H5a. The association of nature experiences with a green energy brand positively influences purchase intention.
H5b. Attitude toward the brand partially mediates the effect of nature experiences on purchase intention.
Fig. 1 presents the proposed theoretical framework of constructs and paths of influence. The empirical study examines the relative strength and significance of the hypothesized effects.

Fig. 1. Influences of perceived benefits and environmental concern on attitude toward green energy brands and purchase intention: theoretical model and structural equation analysis (standardized regression coefficients; NS: non-significant).
Figure options
5. Method and results
5.1. Sample and procedure
For the experimental field study, twelve previously trained interviewers recruited from a final-year undergraduate marketing research course conducted a total of 726 street interviews in six towns and villages in northern Spain. Spanish consumers rate average or somewhat above the European average on many indicators of environmental attitudes, and they are accustomed to the sight of wind turbines and solar power plants (Eurobarometer, 2007). The selection of participants by random street intercept followed quota controls designed to guarantee gender parity and an appropriate age profile for a sample representing potential customers of green energy: female (50%), male (50%); 20–25 years old (20%), 26–35 years old (25%), 36–50 years old (30%), 51–65 years old (20%), and over 65 years old (5%). Gender and age distributions of the sample are acceptably close to the quota. Interviewers showed each participant one of 13 versions of an experimental print advertisement for a fictitious green energy brand, containing one photographic image occupying the majority of the space and several lines of copy (Appendix A). The text is identical in every case, presenting the same information about the brand’s environmental credentials:
eNovis energy offers you clean energy from 100% renewable resources. eNovis energy is generated entirely from sun, wind, water and biofuels. By purchasing the eNovis energy product for a year, a typical household with aver age monthly usage of 1000 kWh per month, could avoid contributing over 17,000 lbs of CO2 into our air — as much as your car makes in almost 20,000 miles of driving.”
Only the advertisements’ visual images varied. Participants were exposed to 12 different natural landscapes ranging from lush green forest to rocky desert, and one cityscape. To avoid interviewer bias and non-random distribution of the experimental advertisements among participants, interviewers selected the advertisement for each interview following a pre-established sequential order. The exposure to different scenery associates the brand with a varying degree of nature experiences, ranging from strongly evoked to none at all in the case of the cityscape, across the sample. After seeing one advertisement, each participant answered a questionnaire assessing attitude toward the brand and its perceived benefits, as well as purchase intention and environmental concern.
5.2. Measurement
The scale for the brand attitude construct consists of one item assessing overall brand evaluation on a ten-point scale anchored by “I like the brand very much” and “I dislike the brand very much” ( [Bergkvist and Rossiter, 2007], [Gresham et al., 1984] and [Yoo and MacInnis, 2005]). Participants rated purchase intention on a traditional 5-point scale measuring the likelihood that they would consider purchasing the brand: “definitely will not buy–probably will not buy–might/might not buy–probably will buy–definitely will buy” ( [Jamieson and Bass, 1989] and [Kalwani and Silk, 1982]). The measurement scales for perceived brand benefits and environmental concern consist of a set of multi-item five-point Likert-scales anchored by “strongly agree” and “strongly disagree”, relating to the statements shown in Table 1, previously tested in several focus group and in-depth interviews with student subjects. Three items measure utilitarian environmental benefits of green energy ( [Johnson and Frank, 2006], [Salmela and Varho, 2006] and [Truffer et al., 2001]). Three items adapted from Nunes and Schokkaert (2003) measure the warm glow construct. The three statements measuring self-expressive benefits draw on research into symbolic consumption (Solomon; 1983) and conspicuous consumption of green products (Griskevicius et al., 2010). With regard to the nature experiences construct, a three-item scale adapted from previous research on human affinity toward and experience of nature ( [Kals et al., 1999] and [Kaplan and Kaplan, 1989]) measures the extent to which a brand evokes feelings similar to those experienced in interaction with actual nature. Finally, three items derived from the New Environmental Paradigm (NEP) scale and several closely related measures assess the environmental concern construct ( [Dunlap and Van Liere, 1978] and [Zimmer et al., 1994]).
5.3. Results
Following validation of the measurement model, structural equation analysis assesses the relationships among latent variables, using AMOS 6 (Fig. 1 and Table 3). The outcomes of this process also confirm an appropriate representation of the underlying data by the proposed factor structure. The analysis reveals a significant positive influence of attitude toward the experimental green energy brand on participants’ intention to purchase (standardized regression coefficient = .36; p < .001). Utilitarian environmental benefits (.16; p = .003), nature experiences (.59; p < .001) and environmental concern (.14; p < .001) significantly affect brand attitude. The perception of utilitarian benefits has a direct influence on purchase intention (.12; p = .001), as well as an indirect effect mediated by brand attitude. The pattern of influences on purchase intention is very similar for environmental concern: a direct effect (.08, p = .036) and an indirect one via its influence on brand attitude. The warm glow construct affects purchase intention only directly (.27; p < .001), and has no significant effect on brand attitude. For the nature experiences dimension, the pattern is reversed: the influence on purchase intention is only indirect, mediated by brand attitude. This construct exerts the strongest of all influences observed in the model, due to its very significant effect on attitude toward the brand (standardized regression coefficient = .59; p < .001). Lastly, the self-expressive benefit construct shows no significant relationship with any of the observed variables.
Table 3. Structural equation analysis: regression coefficients (standardized, unstandardized, p), multiple regression coefficients, model fit.
Addressing the salient influence of the nature experiences construct on brand attitude, the next phase of the analysis assesses the manipulation of the former variable by the variation of the 13 visual images in the experimental advertisements (Table 4). Principal component factor analysis of this construct’s three indicators (Alpha-Cronbach = .82) leads to the extraction of one single factor with eigenvalue > 1.00 that explains 74.4% of the variance and loads on all indicators (factor loadings .82 to .91). ANOVA analysis confirms that mean differences in the ratings of the nature experiences factor (mean = .00; standard deviation = 1.00) between groups of participants exposed to different experimental advertisements are significant (F = 26.01; p < .001). Participants exposed to the image of the mountain creek scenery score highest on the nature experiences dimension (factor score = .37) whereas the urban cityscape and desert landscape rate lowest (factor scores = −.47 and −.69).
The differences in nature experiences across the sample also may affect the pattern of influences of other variables in the proposed model. To establish possible moderating effects of variations in the nature experiences construct, the subsequent analysis, following Baron and Kenny (1986), treats the levels of the moderator as different groups within a multi-group structural equation model (Jöreskog and Sörbom, 1984). For this purpose, the sample is split into three groups with high, medium, and low nature experiences respectively, according to the nature experiences factor ratings within groups exposed to different experimental advertisements, labeled LNE, MNE, and HNE. Table 4 shows that the LNE group, exposed to the desert and city images, scores significantly below mean values (nature experiences factor score = −.58). The MNE group, comprising participants exposed to six different advertisements rating average in terms of nature experiences, has a group mean approximating the mean rating of all advertisements (factor score = .01). The corresponding figures for the HNE group are five advertisements and a factor score of .21. Mean value differences within groups are not significant, but between-group variation is (p < .001).
In the modified model the nature experiences construct constitutes the moderating variable for the multi-group structural equation analysis of the LNE, MNE and HNE sub-samples. Furthermore, lacking significant effects, the self-expressive variable is removed from the model. Fit measures for the multi-group model confirm an appropriate representation of the underlying data by the proposed common factor structure across the three analyzed sub-segments (Table 5). Moreover, the analysis reveals a complex pattern of influences resulting from participants’ varying level of nature experiences. Across all three groups, brand attitude significantly influences purchase intention. In the LNE group, only warm glow affects purchase intention: both directly (standardized regression coefficient = .35) and indirectly, mediated by its influence on brand attitude (.53). The effect of utilitarian benefits on purchase intention in the MNE group is only indirect via its significant relationship with brand attitude (.23), whereas warm glow affects both brand attitude (.32) and purchase intention (.28). In the HNE group, utilitarian benefits and environmental concern significantly influence brand attitude (.26; .21) and purchase intention, both directly (.21; .16) and indirectly, while warm glow exerts direct effects only on brand attitude (.25) at p = .012. Summarizing the results of the multi-group analysis, warm glow has the most significant influence on brand attitude and purchase intention in the LNE condition. Conversely, utilitarian benefits have the strongest effect in the HNE group. Environmental concern affects only HNE participants’ brand attitude and purchase intention.

6. Discussion and implications
6.1. Findings and theoretical implications
Most previous research concerning environmentally friendly consumption behavior focuses on personal factors and personality traits, such as values related to environmental conservation and general concern for the environment. Several studies also address the effect of environmentally sound product attributes delivering utilitarian benefits. Empirical results of this study confirm the influence of consumer’s environmental concern on purchase intention and a partial mediation of this effect by brand attitude (H1a and H1b), consistent with previous research (e.g., [Bang et al., 2000], [Clark et al., 2003] and [Ek, 2005]). Also the perception of environmental utilitarian benefits, such as reduced emissions through the use of renewable energy resources, exerts a significant influence on participants’ purchase intention, partially mediated by attitude (H2a and H2b), confirming assertions in the literature (e.g., [Roe et al., 2001] and [Wüstenhagen and Bilharz, 2006]). With respect to the behavioral effects of psychological benefits related to adoption of green energy, the study confirms the influences of two of the three proposed variables. As hypothesized in the literature (e.g., [Wiser, 1998] and [Wüstenhagen and Bilharz, 2006]), the psychological benefit warm glow arising from contribution to the improvement of the environmental common good increases intention to purchase the experimental green energy brand (H3a). This effect is direct, however, disconfirming partial mediation by brand attitude(H3b) exists. The behavioral effect of warm glow is apparently independent of attitude formation toward a specific brand.
Findings do not support the hypothesized positive influence of self-expressive benefits on brand attitude and purchase intention (H4a and H4b). The mainly private nature of energy use may hinder the perception of such benefits, which presuppose social visibility of consumption behavior. These findings are consistent with Griskevicius et al. (2010), who show that status motives increase desire for green products only when consumed in public, but not in private. Besides, signaling theory posits that the drive for status especially increases the appeal of green products if they are relatively expensive. Electricity, however, is a basic non-expensive commodity, even at a significant price premium for green energy.
A further contribution of this study lies in analyzing the association of green energy brands with natural imagery. Exposed to advertisements showing pleasant natural landscapes, participants report subsequently that the brand evoked feelings similar to those experienced in contact with actual nature. The level of nature experiences varies across the advertisements depicting different landscapes and the one showing a cityscape. Whereas subjects exposed to some of the natural scenes score high on the nature experiences dimension, those exposed to the urban and desert images score significantly lower. The degree of nature experiences evoked positively affects intention to purchase green-branded energy, mediated by its pronounced influence on brand attitude (H5a and H5b). Exploratory findings suggest that manipulation of the level of nature experiences moderates other variables’ relationships. In particular, exposure to natural imagery enhances the influence of environmental concern on purchase intention, possibly by increasing the perceived salience of environmental issues. Environmental utilitarian benefit’s effect also is stronger when participants score high on nature experiences. By contrast, warm glow’s highest overall influence occurs when nature experiences’ level is lowest: that is, following exposure to advertisements in which the visual element is other than pleasant natural landscape.

6.2. Managerial implications
This study’s findings have significant implications for marketers of green energy. Psychological benefits in addition to utilitarian environmental benefits potentially enhance attitude toward green energy brands and increase purchase intention, contributing to renewable energy adoption. For consumers to perceive a significant level of utilitarian benefits, brand communications should supply relevant and sufficiently detailed information. Current energy-labeling schemes are too limited for this purpose. Information supplied should include atmospheric emissions, energy mixes and details of new renewable capacity installed ( [Johnson and Frank, 2006], [Salmela and Varho, 2006] and [Truffer et al., 2001]). To foster the association of warm glow psychological benefits with the brand through appropriate advertising, messages should appeal to the audience’s sense of community, stressing that, when purchasing green-branded energy, they can “feel good while doing good” socially and environmentally (Wiser, 1998). Lastly, the study demonstrates the clear potential of green energy advertisements displaying lush green vegetation and clear water to evoke pronounced psychological brand benefits. Exploratory empirical results on the moderating effects of such advertising induced nature experiences indicate that, at high levels, they could override the warm glow effect, lowering its influence on purchase intention.

6.3. Limitations and future research
The empirical study is primarily of an exploratory nature, focusing on a single green energy brand in an experimental survey and does not measure actual purchasing behavior. However, the questionnaire assesses behavioral intentions which have a stronger relationship with behavior than do attitudes (Fishbein & Ajzen, 1975). Future research should further develop the proposed psychological benefit constructs and confirm findings in a broader setting. Subsequent studies should address on a theoretical level the moderating influences of the nature experiences manipulation. Research on consumers’ mindsets (Kim & Meyers-Levy, 2008) could provide alternative explanations for some of the observed effects.

Consumer attitude toward brand extensions: an integrative model and research propositions
Pengarang       : Sandor Czellar

Abstract     
The paper proposes an integrative model of the antecedents and consequences of brand extension attitude based on the dominant cognitive paradigm. The four key processes of the model are (1) the perception of fit, (2) the formation of primary attitudes toward the extension, (3) the link between extension attitude and marketplace behaviour and (4) the reciprocal effect of brand extension attitude on parent brand/extension category attitude. Moderator and control variables of these processes are identified and classified into three groups: (1) consumer characteristics, (2) marketer-controlled factors and (3) external factors. This integrative model leads to the identification of missing links and variables in past research, resulting in a propositional inventory for future studies. The paper ends with a reflection on the long-term perspectives of scientific inquiry on brand extensions.

1. Introduction
Brand extension is the “use of established brand names to enter new product categories or classes” (Keller & Aaker, 1992, p. 35). The past 15 years have witnessed the development of an important body of empirical evidence on consumer attitude vis-à-vis brand extensions. Systematic research on consumer behaviour toward brand extension was initiated by two seminal North American studies [Aaker & Keller, 1990] and [Boush et al., 1987]. Since, research on the matter has been conducted not only in the United States but also around the world, including countries such as the United Kingdom, France, New Zealand and Taiwan (see [Chen & Chen, 2000], [Holden & Barwise, 1995] and [Sunde & Brodie, 1993]). Many of the effects identified in original studies were later re-investigated by replication studies (see [Glynn & Brodie, 1998] and [Pryor & Brodie, 1998]). The important evolution of the field is reflected by the appearance of the first empirical generalisation based on secondary analysis (Bottomley & Holden, 2001). Brand extension research findings have also been extensively treated from an applied managerial perspective [Kapferer, 1997] and [Keller, 1998].
Recently, Klink and Smith (2001) have warned about a limitation in current research on consumer attitudes toward brand extensions, stating that “in this area, as is often the case during the initial stages of knowledge development, concerns about external validity have taken a back seat to those about internal validity” (Klink & Smith, 2001, p. 326). Indeed, the bulk of research investigates, essentially through experimental designs, the main and interaction effects between a handful of cognitive and affective attitude constructs. Although the studies’ internal validity seems high, their generalisation to real-life decisions and consumption contexts is debatable. Most of them fail to take into account background factors such as individual consumer heterogeneity, marketer-controlled factors and competitive activity, which might exert a significant impact on their generalisation. This article responds to Klink and Smith’s call by offering a guideline for future inquiry on consumer attitudes toward brand extensions in the form of an integrative model and research propositions.
The paper is organised according to a three-step logic, following the structure of previous review studies on other marketing topics (see [Alpert & Kamins, 1994], [Bettman et al., 1998] and [Gatignon & Robertson, 1985]). First, a conceptual model of consumer attitude toward brand extensions is proposed based on the theoretical and empirical developments in the area. Second, the model serves as a guideline for the identification of gaps and underdeveloped areas in past research. Third, research propositions are advanced aiming for the encouragement of empirical inquiry on these underdeveloped areas. The article ends by calling on researchers to adopt alternative conceptual and research paradigms to deepen our understanding of consumer attitude vis-à-vis brand extensions.

2. An integrative model of consumer attitude toward brand extensions
The epistemological stance of research on brand extensions follows the neo-positivist, hypothetical-deductive paradigm of mainstream consumer research [Jacoby et al., 1998] and [Lehmann, 1999]. With notable exceptions, the empirical methods used rely on experimental approaches to identify the main effects, moderators, mediators and control variables in the process of brand extension evaluation. Thus, the bulk of research strives for the development, extension and validation of a general process-based model of the antecedents and consequences of brand extension evaluation.
The following lines offer a description of the evolution of scientific inquiry on consumer attitude toward brand extensions from the point of view of two attitude paradigms: information processing and affect transfer. Two seminal articles laid the ground and, to a large extent, shaped the theoretical basis for empirical research: Aaker and Keller (1990) and Boush et al. (1987). Boush et al. investigated the process of affect transfer from parent brand to the extension [Cohen, 1982] and [Fiske, 1982]. On the other hand, Aaker and Keller focussed on the cognitive process of brand extension evaluation [Fishbein & Ajzen, 1975] and [Fishbein & Middlestadt, 1995]. Loyal to these origins, researchers after 1990 have studied either the information processing side or the affective side of extension evaluation or, more recently, both1(Fig. 1). Fig. 1 depicts an integrative model of consumer behaviour toward brand extensions based on a review of published literature between 1987 and 2001, which is summarised in Table 1. The elements and processes involved in the model are described below.

2.1. Basic process
As for the majority of models of consumer decision-making (see Jacoby, 2002 for a review), the integrative model proposed here is process-based. It is dominated by knowledge and affect transfer processes in the following sequence. Before the appearance of the brand extension in a given product category, consumers already possessed established attitudes both toward the parent brand and the target extension product category. These attitudes are composed of cognitive and affective dimensions [Eagly, 1992], [Fishbein & Ajzen, 1975] and [Fishbein & Middlestadt, 1995]. On the one hand, the cognitive component is brand/category knowledge, defined in terms of the product-related and non-product-related associations linked to a brand/category in long-term consumer memory [Keller, 1993] and [Keller, 1998]. The product-related associations refer to the functional and experiential attributes of the existing products of the brand/category. The non-product-related associations comprise the symbolic benefits stemming from the brand name (such as human personality dimensions, prestige, etc.). On the other hand, the affective component refers to the feelings associated with a brand name or a product category [Boush & Loken, 1991] and [Loken & John, 1993].
When the new extension is launched, consumers evaluate it on the basis of their attitude toward the parent brand and the extension category. If a consumer does not know the parent brand and its products at all, she will evaluate the new extension solely on the basis of her experience with the extension category (Sheinin, 1998). Conversely, if the extension product category is new to her, an attitude toward the extension will be formed only on the basis of her attitude toward the parent brand. If the consumer knows both the parent brand and the extension category, a third effect arises: the perception of fit between the parent brand and the extension category (the components of fit are discussed later in the article). Research has shown that the perception of fit influences extension attitude in two ways. First, it can mediate the transfer of attitude components from the parent brand and extension category to the new extension. Second, fit can moderate the relative influence of brand and category attitude on extension attitude.
Brand extension attitude formation leads to concrete consumer behaviour in the marketplace in terms of intentions, choice and repeat purchase. These experientially based changes in extension attitude give rise to reciprocal effects at different levels. Attitude toward the new extension may affect parent brand attitude in terms of knowledge structure and affect. In a similar vein, attitude to the new extension may influence extension category attitude in terms of knowledge and affect. Both of these reciprocal effects may be moderated by perceived fit.

2.2. Background effects
The basic model depicted in Fig. 1 focalises on the process of extension attitude formation and its effects from the perspective of an individual consumer in isolation. Indeed, past research has essentially investigated consumers’ attitude toward extensions in controlled conditions in a marketplace vacuum. However, in real marketplace conditions, consumers are exposed to a host of information about the extension through different media. Their attitudes toward the extension are sensitive to competitor activity, retailer-level decisions as well as other information sources like press, consumer reports and word-of-mouth. Moreover, the basic model does not account for any heterogeneity in terms of consumer tastes, preferences or consumption situations. In agreement with Klink and Smith (2001), it is argued that the basic model may strongly depend on a series of background factors whose effect should be isolated, investigated and put into perspective with the basic effects of the model. We suggest that these effects be classified into three broad categories: consumer characteristics, marketer-controlled factors as well as external factors. This article investigates a series of variables belonging to these three categories.
Our critical review of research is organised around four themes, which correspond to the major stages of the extension evaluation process. Thus, the article examines successively the processes of:
(1)fit perception,
(2)formation of primary attitudes toward the extension,
(3)link between brand extension attitude and behaviour, and
(4)reciprocal effects of brand extension attitude on parent brand/extension category attitude.
In each of these themes, the state of the art of past theoretical and empirical research is reviewed, leading to the identification of missing links and research gaps. Then, research propositions are formulated that take into account the effect of tentative background factors. The propositions vary in their level of detail depending on the extent of theoretical and empirical evidence on the subject. Thus, some propositions constitute testable hypotheses while others only identify a tentative association between two or more variables. The paper ends with a methodological reflection on the future of brand extension research.

3. Process of fit perception
3.1. Past research
Fig. 2 summarises the current state of research on the process of fit perception. The bold characters and lines in this figure—just as in the subsequent Fig. 3, Fig. 4 and Fig. 5—indicate the topics that have already been subject to scientific inquiry. However, as the reader will see in the following paragraphs, the depth of inquiry on the research topics has been variable. Some relationships are backed with strong empirical evidence while others have only recently been tackled by pioneering work and deserve further replication.

3.1.1. Basic model
Research in this area is based on the categorisation theory of cognitive psychology (Barsalou, 1985). Brands and product categories are conceptualised as cognitive categories in consumer memory [Boush & Loken, 1991] and [Broniarczyk & Alba, 1994]. A brand extension in a new product category is viewed as a new instance that can be more or less similar to the brand and its existing products. Perceived similarity, also called perceived fit, is characterised by the number of shared associations between the extension product category and the brand. Researchers have identified two dimensions of the fit construct [Bhat & Reddy, 2001] and [Park et al., 1991]. The first is product category fit, which refers to the perceived similarity between the extension category and the existing product categories of the parent brand. The second is brand-level fit, referring to the match between the specific image of the brand and the extension product category. To illustrate these two aspects of fit, consider Marlboro launching a ball-pen. The perceived fit between Marlboro and the ball-pen category will be composed of a category-level fit (the shared product attributes between cigarettes and ball-pens) and a brand-level fit (e.g. the match between Marlboro’s rough, Western brand image and the image the consumer holds about the ball-pen category).

3.1.2. Background factors
Research on the background factors of this basic fit perception process is relatively scarce. It deals mainly with the effect of consumer mood and advertising on fit perceptions. Thus, recent research shows that positive consumer mood improves fit perceptions for moderately far extensions (Barone et al., 2000). To our knowledge, however, research on other consumer-level factors has not yet been undertaken. Advertising can be used in several ways to directly improve consumers’ fit perceptions. Through increased exposure, it can facilitate information retrieval processes and thus improve fit perceptions [Klink & Smith, 2001] and [Lane, 2000]. By manipulating the informative content of ads through priming and distancing techniques, marketers can also enhance fit perceptions [Boush, 1993], [Kim et al., 2001] and [Pryor & Brodie, 1998]. However, marketers can also use marketing-mix variables other than advertising to improve fit perceptions. These, as well as other possible background factors such as competitor and distributor activity, have not yet been studied. The following lines indicate several paths for future research in these underdeveloped areas.

3.2. Research propositions
This section investigates the effect of a series of consumer, marketer-controlled and external factors that can influence the role of parent brand and category knowledge on fit perception (see the plain characters and lines

3.2.1. Consumer characteristics
Research on consumer memory shows that expertise with a specific product category leads to more and more elaborate and complex knowledge structures (Alba & Hutchinson, 1987). It also appears that brand ratings by expert consumers are based on concrete product attributes, whereas the brand ratings of novices stem from more general impressions about the brand (Dillon, Madden, Kirmani, & Mukherjee, 2001). On the other hand, less experience leads to less concrete category and product knowledge and more reliance on symbolic associations and general impressions about the brand [Braun & Wicklund, 1989] and [Dillon et al., 2001]. Therefore, the following statement is advanced:
P 1. Higher consumer expertise leads to the greater relative role of product-related brand associations vs. non-product-related brand associations in fit perception.
According to self-monitoring theory, people differ substantially in the way they regulate their self in public situations [Gangestad & Snyder, 2000] and [Snyder, 1974]. Low self-monitors tend to project a stable self in diverse settings of social interaction. Their behaviour is guided more by inner psychological factors than social influences. High self-monitors, on the other hand, exert more expressive control over their social behaviour and tend to adapt their appearance and acts to specific situations. Empirical research has shown that high self-monitors respond more favourably to status-oriented advertising claims [DeBono, 1987] and [DeBono & Harnish, 1988] and prefer brands in congruence with social situations [Aaker, 1999] and [Hogg et al., 2000]. Note that the role of non-product-related brand associations is primarily symbolic and self-expressive [Keller, 1993] and [Solomon, 1983]. Therefore, during the entire process of brand extension evaluation, high self-monitors can be expected to confer more importance to these associations than their low self-monitor counterparts. In this section, it is argued specifically that high self-monitors perceive fit more on the basis of non-product-related associations than low self-monitors. In formal terms:
P 2. Non-product-related associations have a greater effect on fit perceptions for high self-monitors than for low self-monitors.

3.2.2. Marketing strategy
Independently of the product category, Park, Jaworski, and McInnis (1986) defined three broad types of brand positioning derived from basic consumer needs. Functional needs stem from motivations to solve externally generated concrete problems and triggers a search for products that provide solutions to these problems. Brands with a functional positioning offer these solutions; their value is dependent on satisfaction after use. Experiential (also called hedonic) needs correspond to a desire for sensory pleasure and stimulation. Therefore, brands with an experiential positioning emphasise satisfaction-in-use. A third category of needs, that of symbolic needs, concerns the individual’s desire for self-identity creation, maintenance and enhancement. Thus, brands with a symbolic positioning enable their consumers to be associated with a desired group, role or self-concept. Park et al. (1991) showed empirically that symbolic brands, characterised by the dominance of non-product-related associations, are easier to stretch to more dissimilar product categories than functional brands. In the same spirit, we argue here that fit perceptions of symbolic brands are influenced mainly by non-product-related associations; the fit of a functional or experiential brand with the extension category, on the other hand, will be determined basically by product-related associations.
Furthermore, the consumer’s personal characteristics may interact with the effect of brand positioning. It is argued that the effectiveness of the brand’s positioning strategy depends on the targeted consumers’ characteristics in terms of expertise, cognitive capacity and self-monitoring. Expert consumers possess elaborate and complex knowledge structures about a given product category (Alba & Hutchinson, 1987). It is therefore relatively more difficult to alter these structures by new pieces of information than the less elaborate knowledge structures of novices. Human cognitive capacity is strongly linked to age. Research shows that the cognitive capacity of the elderly, especially over 65, declines progressively [Chasseigne et al., 1997] and [Lambert-Pandraud & Laurent, 2002]. For these people, the learning of new information becomes difficult and they tend to rely on existing information in long-term memory rather than on active short-term memory to make judgements and decisions (Salthouse, 1991). We therefore expect elderly people to be less sensitive to new information on the product’s positioning than younger people. In addition, it is argued that the self-monitoring level of the consumer may also impact on her/his sensitivity to external information because high self-monitors are more influenced by external contexts than low self-monitors. Hence,
P 3. For brands with a functional or experiential positioning, providing product-related information influences fit perceptions more than providing non-product-related information. Expert consumers, low self-monitors and elderly people are less sensitive to these actions than novice consumers, high self-monitors and younger people.
P 4. For brands with a symbolic brand positioning, providing non-product-related information influences fit perceptions more than providing product-related information. Novice consumers, high self-monitors and younger people are more sensitive to these actions than expert consumers, low self-monitors and elderly people.

3.2.3. External information
Competitor activity refers to all the marketing actions that competing brands—already present in the extension product category or newly entering it—might undertake. It is reasonable to assume that the only effect that might jeopardise the new extension is the one emanating from competitors offering a similar positioning. Thus, if Marlboro decided to launch a low-end, mass-market deodorant, consumer attitudes toward this extension would not be influenced by the marketing activity of luxury brands like Chanel or Dior but rather by cheaper brands like Denim or Nivea. It is thus assumed that direct competitors provide the consumer with rather similar product-related or non-product-related messages, similar to those of the company’s new extension. By processing this external information, consumer fit perceptions may be altered. For the same reasons as in the case of marketer-controlled factors, the effectiveness of competitive activity depends on the targeted consumers’ characteristics in terms of expertise, self-monitoring and age. Another source of information for consumers is the point-of-purchase, which is controlled by the brand’s current distributors. The distributors’ marketing activities may also provide potential consumers with information that is relevant for their judgements about the fit between the brand and the extension category. Other external information such as word-of-mouth may also have an impact on fit perceptions. It is therefore advanced that:
P 5. Competitor marketplace activity, distributor activity and other external information directly affect the perceived fit between the brand and the extension. Novice consumers, high self-monitors and younger people are more sensitive to these actions than expert consumers, low self-monitors and elderly people.
Preliminary empirical evidence suggests that product categorisation processes are context dependent (Wanke, 1999). That is, depending on the context, the same objects can be categorised in different sets by the same consumer. The importance of this phenomenon for brand extensions may depend on the self-monitoring style of the consumer. A basic tenet of self-monitoring theory is that low self-monitors show a more stable behaviour across contexts than do high self-monitors (Snyder, 1974). It is argued that low self-monitors, whose social behaviour is relatively invariant, will tend to keep the same categorisation sets across consumption situations. High self-monitors, on the other hand, will adapt their categorisation schemas to the social context. For example, they may perceive a new Marlboro deodorant as having a higher fit in private consumption situations than in public ones. By opposition, the fit perceptions of low self-monitors are expected to be more invariant across consumption situations. We can therefore expect that:
P 6. Fit perceptions vary more across situations for high self-monitors than for low self-monitors.

4. Extension evaluation process
4.1. Past research
4.1.1. Basic model
Bold lines and characters in Fig. 3 depict the concepts and relations referring to the formation of brand extension attitude that have already been investigated to some extent in past research. As mentioned in Introduction, cognitive processing and affect transfer theories dominate current research on brand extension attitudes. Research shows that, all else being equal, there is a direct knowledge/affect transfer from the parent brand to the extension (see the references in Table 1). For example, the perceived high quality of the parent brand results in positive extension evaluations (Aaker & Keller, 1990). Similarly, positive feelings are directly transferred from the parent brand to the extension (Bhat & Reddy, 2001).
The concept of fit is central in past research on brand extension attitudes. One of the most frequently studied topics is the direct effect of fit on brand extension attitude (see the references in Table 1). The general conclusion from these studies is that the higher the perceived fit, the more positive the consumer’s attitude toward the extension. The vast majority of studies have examined the direct effect of category-level fit; only few studies are devoted to the effect of brand fit (e.g. [Bhat & Reddy, 1997], [Bhat & Reddy, 2001] and [Park et al., 1991]).
Apart from being modelled as a direct effect, researchers have also considered the moderating role of fit on brand knowledge and affect transfer. The level of fit determines the ease of transfer of positive knowledge and affect from the parent brand to the extension. Researchers have extensively studied the moderating role of fit on several aspects of knowledge transfer: general quality, technological level, specific product attributes (see references in Table 1). Also, the higher the perceived fit, the higher the affect transfer from the parent brand to the extension (e.g. [Broniarczyk & Alba, 1994] and [Gürhan-Canli & Maheswaran, 1998]).

4.1.2. Background factors
Past research has included a series of consumer-specific moderators in the basic effects model, namely motivation, expertise, implicit personality theory, innovativeness and mood (see Table 1). Strong empirical evidence supports the moderating effect of motivation [Gürhan-Canli & Maheswaran, 1998] and [Nijssen et al., 1995]. Specifically, Gürhan-Canli and Maheswaran (1998) show that under high-motivation conditions, elaborate cognitive processing is expected: consumers consider every piece of information about the extension piecemeal. Conversely, under low-motivation conditions, cognitive processing is less elaborate and more categorical; in this case, more affect transfer is expected from the parent brand to the extension. Also, in high-motivation conditions, perceived fit has less impact than in low-motivation conditions. Consumer expertise is also shown to moderate the effect of product-related brand associations, brand affect and fit on brand extension attitude (Broniarczyk & Alba, 1994). According to the findings, expert evaluations are based on the processing of product-related associations, whereas novices tend to evaluate the extension more on the basis of brand affect and fit. Consumer innovativeness is another factor that influences the basic effects of the extension evaluation model (Klink & Smith, 2001). Highly innovative early adopters are less sensitive to risk; fit therefore plays a lesser role in their extension evaluations than in those of late adopters.
Flaherty and Pappas (2000) show the effect of a psychological evaluation process (implicit personality theory) on affect transfer. Research in personality psychology shows that entity theorists easily form global judgements based on prior trait information, whereas incremental theorists tend to make more conditional judgements upon situational cues (Hong, Chiu, Dweck, & Sacks, 1997). Flaherty and Pappas show that affect transfer from brand to extension occurs more easily for entity theorists, who are more sensitive to existing parent brand beliefs than for incremental theorists, whose information processing is more elaborate.
Still on the consumer-specific side, recent research suggests that positive consumer mood enhances attitude transfer from brand to extension (Barone et al., 2000). The effect of mood is strongest in moderately far (vs. near or far) extensions in terms of fit.
Ample empirical evidence shows that elements of advertising strategy such as information type, information amount, exposure, as well as techniques such as priming and distancing, can affect consumer attitude toward extensions (see references in Table 1). That is, by providing and manipulating the information about the new extension, marketers can directly improve consumer attitude, thereby reducing the importance of cognitive processing, affect transfer and fit in consumer evaluations. This issue is particularly relevant in situations where the proposed extension is highly incongruent with the existing products of the brand or when consumer attitude toward the parent brand is rather negative.
In sum, research has mainly investigated the cognitive and affective processes of attitude formation toward extensions at an individual level. To our knowledge, only Sheinin (1998) has investigated the effect of knowledge transfer from the extension product category to the new brand extension. Another topic scarcely investigated in the basic model is the role of non-product-related associations in attitude formation. Moreover, future research should broaden the scope of the model by an in-depth examination of its sensitivity to further consumer-level, marketer-controlled and external factors. The research propositions advanced below serve to foster such research efforts.

4.2. Research propositions
The research propositions investigated below refer to the concepts and relations in plain characters and lines

4.2.1. Consumer characteristics
Previous research has established the role of consumer expertise in the processing of the product-related associations of the parent brand (Broniarczyk & Alba, 1994). However, more and more brands are positioned on axes like personality traits and user imagery, which are relatively independent of the product features. These non-product-related associations are typically less experience-based than the concrete, product-related features of the brand. Novice consumers are therefore expected more to rely on non-product-related associations in extension evaluation than their experimented counterparts. Furthermore, this effect can be generalised to category-related associations, by assuming that these associations follow the same pattern of processing as that of brand-related associations.
P 7. Higher consumer expertise with the parent brand/extension category leads to a greater transfer of product-related associations vs. non-product-related associations from parent brand/extension category to the brand extension.
Broniarczyk and Alba (1994) have also shown that experience with the parent brand leads to a lesser reliance on fit in brand extension attitude formation. At the same time, affect transfer declines with consumer experience to the benefit of cognitive effects. The next two propositions generalise these conclusions to the experience with the extension product category, too.
P 8. The higher the consumer’s expertise with the parent brand/extension category, the lesser the impact of perceived fit on extension attitude.
P 9. Knowledge transfer increases with consumer expertise with the parent brand/extension category, whereas affect transfer decreases with consumer expertise.
As explained previously, research in social psychology shows that high self-monitors attach more importance to the self-expressive and social meaning of products than low self-monitors [Aaker, 1999], [DeBono, 1987] and [DeBono & Harnish, 1988]. In this respect, non-product-related associations, which are related to the self- and value-expressive benefits of the brand, are expected to play a greater role in brand extension knowledge for high self-monitors than for low self-monitors. Hence,
P 10. Non-product-related associations of the parent brand/extension category are transferred more easily to the extension for high self-monitors than for low self-monitors.

4.2.2. Marketing strategy
Previous research has mainly examined the effect of advertising content, amount and frequency on brand extension evaluations. The long-term positioning of the brand, supported by all the elements of the marketing-mix, is likely to play an important role in brand extension evaluation. Consumers tend to have more product-based expectations about brands with functional or experiential positioning. Symbolic brands, on the other hand, are expected to convey more non-product-related associations. It is also proposed that consumer sensitivity to these marketing actions is dependent upon expertise, self-monitoring and age.
P 11. For brands with a functional or experiential positioning, product-related information is processed more than non-product-related information in brand extension evaluation. Novice consumers, high self-monitors and younger people are less sensitive to these actions than expert consumers, low self-monitors and elderly people.
P 12. For brands with a symbolic positioning, non-product-related information is processed more than product-related information in brand extension evaluation. Novice consumers, high self-monitors and younger people are less sensitive to these actions than expert consumers, low self-monitors and elderly people.

4.2.3. External information
 As mentioned in the previous section, competitor activity refers to all the marketing actions that competing brands may envisage in the extension category. It was assumed that only competitors offering a similar positioning might endanger the new extension. Consumers process the information communicated by competitors, which leads to changes in their knowledge structures vis-à-vis the extension. By the same logic, point-of-purchase distribution decisions may also provide consumers with further information about the extension and its competitors. The effectiveness of competitive and distributor activity as well as any other external information related to the extension may depend on the targeted consumers’ characteristics in terms of expertise, self-monitoring and age.
P 13. Competitor marketplace activity, distributor activity and other external information directly affect brand extension knowledge. However, novice consumers, high self-monitors and younger people are more sensitive to these actions than expert consumers, low self-monitors and elderly people.

5. Link between extension attitude and marketplace behaviour
5.1. Past research
Past research has devoted relatively little attention to the study of the direct link between brand extension attitude and marketplace behaviour (see Table 1 and the bold lines and characters.
Experimental studies show that positive affect leads to higher purchase intentions for the extension [Bhat & Reddy, 2001] and [Lane, 2000]. Empirical evidence on the role of extension knowledge structures in the extension attitude–behaviour relationship is scarce and rather indirect. Sullivan (1992), based on secondary sources, used aggregate brand-level data to show that brand extensions introduced early in the life cycle of a product category did not perform as well as extensions introduced at later stages. Reddy et al. (1994) used aggregate secondary data and expert judgment to demonstrate that the brand’s symbolic associations have a positive impact on the extension’s market share. Swaminathan et al. (2001) illustrated with panel data that prior experience with the parent brand leads to a higher probability of extension trial. The results of these three studies seem to suggest that the more elaborate the consumer’s brand extension knowledge, the more likely she is to purchase the extension.
However, to our knowledge, no academic studies have systematically investigated the link between extension attitude and the marketplace behaviour of the individual consumer. Consider the following intriguing questions: What is the relative role of extension knowledge and extension affect on intentions, choices and repeat purchase? Which part of brand knowledge (product-related or non-product-related associations) is more important in consumer decision-making? How does concrete experience with the extension product impact extension attitude? What are the consumer characteristics that shape these effects? The following lines offer a series of propositions to guide future research on these topics.

5.2. Research propositions
In the present context, this section focalises only on the link between extension attitude and behaviour; other internal and external factors directly affecting consumer market behaviour vis-à-vis the extension are beyond the scope of this article (e.g. socio-economic status, deal proneness, competitive activity, fashion trends, etc.). The research propositions investigated below refer to the concepts and relations in plain characters and lines in Fig. 4. The framework is based on the assumption that, depending on the consumer’s characteristics, specific attitude components will guide the relationship between extension attitude and marketplace behaviour. That is, specific extension attitude components (shaped by parent brand and category knowledge, brand marketing activity and other information sources) exert an impact on consumer marketplace behaviour. Later, positive and negative experiences with the extension will impact on specific extension attitude components, depending on personal characteristics.
The research propositions concerning the role of consumer characteristics are formulated in the spirit of our arguments in the previous sections. Just as consumer characteristics can moderate the basic effects of fit perception and extension attitude, it is proposed here that they can also moderate the link between extension attitude and behaviour. Specifically, it is argued that consumer expertise leads to the dominance of the concrete product knowledge about the parent brand and the extension. For experts, purchase intentions and choice, as well as repeat purchase, will be guided by detailed product-related knowledge about the brand. Their experiences with the extension will also focalise on product-related aspects, which will further contribute to their extensive knowledge about the product. Novices, on the other hand, know relatively little about the concrete product characteristics of the parent brand and its proposed extension. Their purchase intentions, choice and repeat purchase will be based more on non-product-related associations and affect. Through experience, they can progressively acquire more knowledge about the product-related aspects of the brand extension. It is therefore proposed that:
P 14. Consumer expertise moderates the relationship between extension attitude and marketplace behaviour. Specifically, non-product-related associations and affect play a greater role on this relationship for novices than for experts.
The previous sections dealt with the moderating role of self-monitoring on fit perceptions and extension attitude. It can be argued that self-monitoring can affect the link between extension attitude and behaviour, too. High self-monitors are more sensitive to the social symbolic meaning of brands (prestige, status, personality) than low self-monitors [DeBono, 1987] and [Hogg et al., 2000]. It is likely that for high self-monitors, considerations about the non-product-related associations will play a greater role in purchase decisions than product-related associations. Also, the performance of the brand will be judged more on the basis of its symbolic value than on its experiential or functional value, contributing mainly to the non-product-related knowledge about the brand. The following proposition is therefore advanced:
P 15. Consumer self-monitoring moderates the relationship between extension attitude and marketplace behaviour. Specifically, non-product-related associations play a greater role on this relationship for high self-monitors than for low self-monitors.
The third consumer-specific variable examined here is age. As mentioned before, the progressive decline of the cognitive abilities of elderly people has been documented in psychology (Chasseigne et al., 1997). Elderly people tend to extensively rely on their long-term memory, whereas fluid, working memory is more limited in its accessibility and acceptation of new information. It is argued that, with the increase of age, knowledge will play a lesser role, whereas affect will play a greater role in consumer decision-making. That is, older people will judge products more on affective bases than on cognitive bases.
P 16. Consumer age moderates the relationship between extension attitude and behaviour. Specifically, affect plays a greater role in this relationship for elderly people than for younger people.

6. Reciprocal effects of brand extension attitude on parent brand and extension category attitude
As discussed in the previous sections, through diverse information sources (marketing actions, competitor information, distributor activity, etc.) and/or direct experience, consumers form either a positive or a negative attitude toward the extension. Attitude valence may take a cognitive form in terms of favourable/unfavourable associations or an affective form of liking/disliking the extension. Attitude toward the extension may alter the consumer’s original attitude toward the parent brand or the extension category. This phenomenon is known in the literature under the term “reciprocal effect” of brand extension, which may lead to dilution/enhancement of the original brand/category attitude [John et al., 1998], [Loken & John, 1993] and [Romeo, 1991].

6.1. Basic model
The direct effects of reciprocal knowledge transfer (e.g. [John et al., 1998], [Loken & John, 1993] and [Milberg et al., 1997]) and affect transfer (e.g. [Keller & Aaker, 1992] and [Romeo, 1991]) have been extensively documented in past research. However, research focussed mainly on knowledge transfer effects (see Fig. 5 and Table 1). It is evidenced that product-related negative associations with the extension dilute product-related parent brand associations, whereas non-product-related extension associations dilute non-product-related parent brand associations (Chen & Chen, 2000). In addition, general product-related parent brand associations (e.g. quality) are more difficult to alter than specific product-related associations (e.g. taste) [Keller & Aaker, 1992] and [Loken & John, 1993]. Information accessibility also plays a role in knowledge transfer from extension to the brand (Ahluwalia & Gürhan-Canli, 2000). Less accessible knowledge structures concerning the extension, stocked in long-term memory, have less effect on the parent brand than highly accessible new information.
The moderating role of fit on reciprocal effects has also received sustained attention (e.g. [Ahluwalia & Gürhan-Canli, 2000], [Boush & Loken, 1991] and [Morrin, 1999]). Increased fit generally leads to increased knowledge and affect transfer from the extension to the parent brand. However, it seems that dilution effects in terms of parent brand knowledge and affect are more likely to occur for close extensions, whereas enhancement effects are more likely to occur for far extensions (Ahluwalia & Gürhan-Canli, 2000).

6.2. Background factors
Consumer characteristics as moderators have also been included in the basic model, namely, motivation and expertise. In high-motivation conditions, consumers process every piece of new information in detail. In these situations, associations with the extension may alter the parent brand’s association network regardless of fit (Gürhan-Canli & Maheswaran, 1998). In low-motivation conditions, however, this effect is dependent on fit because less similar extensions will be quickly categorised as atypical instances and will not alter parent brand knowledge. Consumer expertise also moderates the dilution/enhancement effect of brand extensions on parent brand knowledge. Experts possess strong brand-related memory structures, which are difficult to affect by either negative or positive extension information [John et al., 1998] and [Morrin, 1999]. More specifically, extensions influence more the product-related associations of unfamiliar parent brands than those of familiar parent brands (Sheinin, 2000).
At the level of firm-related background factors, Morrin (1999) shows that exposure to extension advertising might strengthen existing parent brand knowledge structures. Retailer-level decisions, on the other hand, can weaken parent brand knowledge. Thus, the concentration of the new extension with competing brands on shelf space tends to confuse consumers and makes extension advertising less effective.

6.3. Research propositions
In sum, past research has mainly investigated the reciprocal effect of extension attitude on the parent brand; the effect on the consumer’s attitude toward the extension product category has scarcely been investigated. Moreover, past studies have essentially dealt with the transfer of product-related associations and less with the transfer of non-product-related associations or affect from the extension to the parent brand. A series of propositions are advanced aiming at filling these gaps. The research propositions refer to the concepts and relations in plain lines in Fig. 5. The reasoning behind them is quite similar to that put forward in the section on the primary attitude transfer from parent brand/category to the extension.

6.3.1. Consumer characteristics
The first proposition refers to consumer expertise. It was mentioned before that expert consumers possess a more elaborate product-related knowledge structure about the parent brand and the product category than novice consumers. Through external information search and concrete product experience with the new brand extension, their knowledge concerning the new extension will also be mainly product-related. Concrete product-related associations are therefore more likely to be transferred from the extension to the parent brand/product category than non-product-related associations, which are more abstract and symbolic.
P 17. Higher consumer expertise leads to the greater relative transfer of product-related associations vs. non-product-related associations from the extension to the brand/category.
The next proposition posits that self-monitoring also moderates the process of reciprocal attitude transfer from the extension to the parent brand and the product category. High self-monitors value more the symbolic attributes of the extension than low self-monitors. Their attitude toward the extension is shaped essentially by the extension’s capacity to provide self-expressive or social symbolic benefits. It can be hypothesized that their extension attitude will only enhance parent brand/extension category attitude inasmuch as it provides non-product-related benefits. It is hence expected that:
P 18. Non-product-related associations are more easily transferred from the extension to the brand/category for high self-monitors than for low self-monitors.
In the previous sections, the tentative effect of age on brand extension evaluations was investigated. Based on prior literature in cognitive psychology, it was argued that for elderly consumers, learning and processing of information about a new extension is difficult. Their extension attitude will be more of a matter of affect transfer from the parent brand and the product category. Following this reasoning, the next proposition states that affect transfer dominates the reciprocal effect of extension attitude on parent brand/product category.
P 19. Affect transfer from extension to parent brand/extension category increases whereas knowledge transfer declines for elderly consumers.

6.3.2. Marketing strategy and external information
The last two propositions investigate the effect of extension marketing strategy and external information sources on reciprocal effects. As mentioned earlier, the effect of information amount and exposure length has already been studied in this respect (Morrin, 1999). Here, the focus is on the effect of information type. Any information about the brand extension might alter not only the consumer’s attitude toward the extension but also directly her/his attitude toward the parent brand and the product category. The hypothesis is that the reciprocal effect is dependent upon the positioning of the parent brand or the product category (functional, experiential, symbolic). For brands/categories with a functional or experiential positioning, positive (negative) product-related information will weigh more than non-product-related information. Further, it is proposed that the size of this effect is dependent upon the characteristics of the consumer in terms of expertise, self-monitoring and age.

P 20. For brands/categories dominated by experiential/functional positioning, consumer knowledge about the brand/category is more sensitive to product-related extension information than to non-product-related extension information. This effect is stronger for novice consumers, low self-monitors and younger people than for expert consumers, high self-monitors and elderly people.

P 21. For brands/categories dominated by symbolic positioning, consumer knowledge about the brand/category is more sensitive to non-product-related extension information than to product-related extension information. This effect is stronger for novice consumers, high self-monitors and younger people than for expert consumers, low self-monitors and elderly people.

7. Future research directions
The first 15 years of research on consumer attitude toward brand extensions have seen the development of an elaborate process-based model grounded in cognitive psychology. Ample empirical evidence attests the main effects in the model and interactions between them. Researchers have also started to investigate consumer, marketer-controlled as well as external factors that moderate the relations in the model. This critical review has identified a certain number of gaps, underdeveloped aspects and possible extensions of this essentially cognitive model. To advance knowledge, future experimental studies can be guided by the empirical investigation of the propositional inventory presented here. Also, new consumer-level, marketer-controlled and external factors could be added and tested in the proposed integrative model.
This brings me to conclude the article with a critical note on the current research paradigm. On the methodological side, experimental studies have largely dominated scientific inquiry on brand extensions. Yet such a confinement may jeopardise the long-term perspectives for the development of a field. Eagly (1992) warns that “if investigators look to only the most obviously relevant research, not only do they miss many potentially useful theoretical ideas, but also they allow their theories to be seriously limited by the constraints of their research paradigms, which often allow only certain processes to be manifest. Theory encapsulated within an experimental paradigm is thus limited in scope” (Eagly, 1992, p. 704). In agreement with Eagly, this article encourages researchers to broaden their perspectives in future investigations of consumer attitudes toward brand extensions.
In addition to experimental research, large-scale studies based on longitudinal data can be employed to extend the empirical evidence of the cognitive model outlined in this article. Specifically, they can shed light on the influence of increasing category familiarity and expertise on brand extension evaluations. Periodically repeated surveys with the same consumers (e.g. panel data) are particularly useful for the investigation of the effect of consumer expertise on fit perception P 1, P 3, P 4 and P 5, primary extension attitude P 7, P 8, P 9, P 11, P 12 and P 13, the attitude–behaviour link (Proposition 14) and reciprocal effects P 17, P 20 and P 21. Specifically, such longitudinal studies might uncover a possible nonlinear effect of an increase in consumer expertise on extension attitude formation. For example, could it be that the effect of expertise increase follows an S-shaped curve, whereby its effect is weak at early periods, then grows progressively and becomes roughly linear, while decreases down again at later stages?
Besides survey-type research, qualitative methodologies may also prove useful in broadening the scope of our knowledge about brand extension evaluations. In-depth interviews with consumers may extend research within the current cognitive theoretical paradigm, essentially through a deeper understanding of the role of non-product-related associations in fit perception P 1, P 2, P 3 and P 4 and extension attitude formation P 7, P 10, P 11 and P 12. Quantitative studies have failed to identify a typology of possible non-product-related associations that might shape extension evaluation, such as user imagery and brand personality dimensions (Keller, 1998). Indeed, consider these questions: Are these concepts applicable to both brands and product categories? If so, what is the relative role of user imagery and brand personality in brand-level fit perception? Is brand/category user imagery or brand/category personality more important in the consumer’s evaluation of a brand extension? Under which consumer-level, marketer-controlled and external background factors is the former more salient, and therefore more easily transferred, than the latter? In this respect, researchers may use the critical incident method, advocated by Lincoln and Guba (1985): the analysis of detailed consumer narratives about recent or older brand extension experiences may elucidate answers to the questions raised.
In-depth interviews with brand managers and case studies offer other promising paths for future research. They can enhance our knowledge about the influence of managerial decisions, competitive activity and other external factors on the extension evaluation process. In particular, these information sources may help us to better understand which specific elements of the marketing mix are more effective than others in increasing fit perceptions P 3, P 4 and P 5 and favourable extension knowledge P 11, P 12 and P 13. In addition, insights from managers may also show if the same marketing mix elements have the same weight on extension evaluations, whether the company or its competitors use them.
Last but not least, in the spirit of Fournier and Mick (1999), qualitative consumer research can be used to challenge the dominant cognitive paradigm and test alternative theoretical perspectives on brand extensions. One of such perspectives is the emergent relationship paradigm, which sees brands as humanlike partners in types of relationships with the consumer that vary both in quality and intensity (Fournier, 1998). From this standpoint, brand extension can be conceptualised as an important type of brand behaviour that affects the relationship developed between the consumer and the brand. What is the impact of a brand extension announcement on the evolution of brand relationship quality? How do concrete experiences with the extension affect the relationship? Also, which relationship types are conducive to which types of brand extension in terms of positioning and level of fit ? Indeed, do consumers with stronger brand relationships tolerate more dissimilar extensions than consumers with looser brand ties? Or is it the opposite, in the sense that some strong relationships may lead to a “possessive” brand attitude, whereby the consumer becomes less favourable to drastic changes in the brand offer? Future inquiry is warranted to answer these interesting questions. Indeed, such an enterprise is likely to move us from the current linear, individual and predominantly process-based view of brand extension evaluation toward a holistic understanding of the relation between a consumer and a brand extension.

The role of the social-identity function of attitudes in consumer innovativeness and opinion leadership
Pengarang       : Rajdeep Grewal

Abstract
Attitudes serving the social-identity function relate nonsocial objects (e.g., products) to social objects (e.g., people). As new products tend to be more exciting than old, familiar products, the authors suggest that these attitudes influence innovativeness and opinion leadership. Based on recent research on attitude functions and adoption of consumer innovations, this research examines the relationship between the social-identity function, innovativeness, and opinion leadership, in addition to expertise and involvement; the two traditional antecedents of innovativeness and opinion leadership. The results across two product categories show that social-identity function exerts a strong impact on innovativeness and opinion leadership.

1. Introduction
The recent resurgence of interest in the motivational underpinnings of cognitive constructs (Kruglanski, 1996) has stimulated interest in issues relating to the psychological needs served by attitudes (cf., Greenwald, 1989). One related and important stream of research concerns attitude functions (cf., Shavitt, 1992). Functional theories of attitudes (cf., [Katz and Stotland, 1959] and [Sarnoff and Katz, 1954]) ask the question: Why do people hold the attitudes they do? Development of new methods to measure attitude functions helped overcome the operational difficulties that have plagued research on attitude functions (Herek, 1987). We suggest that attitude functions may play an important role in the adoption and the diffusion of consumer innovations (Gatignon & Robertson, 1985). The success of new consumer products rests on the construct of innovativeness, which introduces the product (innovation) to the social system, and opinion leadership, which provides social legitimacy to the innovation. Thus, it is important to develop an understanding both innovativeness and opinion leadership.

2. Theoretical background and research hypothesis
Functional theorists classify attitudes according to the functional needs that they meet [Snyder and DeBono, 1989] and [Shavitt, 1989a]. General classifications categorize attitude functions as utilitarian, ego-defensive, knowledge, value-expressive, and social-adjustive (Smith, Bruner & White, 1956). One important objective of consumer psychology is to be able to persuade consumers and change their attitudes. According to the functional theorist, the prerequisite for changing an attitude is to determine the psychological need served by the attitude (Shavitt, 1989a).

This role of attitude functions opens a plethora of opportunities to investigate persuasion processes [Johar and Sirgy, 1991], [Park et al., 1991] and [Shavitt, 1992]. Indeed, most applications of attitude functions in consumer behavior investigate the persuasiveness of advertising messages (cf., [Park et al., 1986] and [Snyder and DeBono, 1985]. For example, Shavitt and Lowrey (1992) investigate the impact of product-related and audience-related factors on the persuasiveness of value-expressive and utilitarian advertising messages. Other applications examine objects serving different attitude functions (cf., [Shavitt and Fazio, 1991] and [Sirgy et al., 1991]. For instance, in their investigation of brand extensions, Park et al. (1991) explore functional (utilitarian function) and prestige (social-adjustive function) related brand names.
Our research studies the social-identity attitude function within the context of consumer innovations. Shavitt (1990) associated the self-expression (value-expressive) and social interaction (social-adjustive) role of attitudes to symbolize the social-identity function. Social-identity function defines a larger symbolic category of attitudes [Abelson, 1982] and [Sears and McCohanay, 1973] and mediates the self and the other interactions (Smith et al., 1956). This function discriminates between individuals who prefer to display their true inner self and those for whom attitude helps to harmonize into important social situations (Snyder & DeBono, 1989).
Products help one achieve private and public identity goals [Abelson and Prentice, 1989] and [Shavitt, 1989a]. Product perception is a social process if the perceiver perceives the product, either consciously or unconsciously, as a social entity. Products also help in defining self and to maintain important self-definitions (cf., Belk, 1988) and, often provide people with a characteristic they feel they lack. In fact, people exhibit self-enhancing biases when making judgments about product they own, the mere ownership effect (Beggan, 1992). The products that perform the social-identity function: (1) mediate the self and other relationships, (2) instill pride in a person as they help the person to fit in desired social settings, and (3) determine the centrality of the object to the person.
Innovations, because they are generally exciting and trend setting, become primary candidates for eliciting the social-identity function. For example, if cars are important for portraying self-image for Sam, then Sam is highly likely to own a state-of-art model of car. Thus, it is our contention that the social-identity function of attitude plays an important role in the adoption and the diffusion of innovations. To gain a better understanding of this phenomenon, we investigate the relationship between the social-identity function and the parameters of consumer innovations: innovativeness and opinion leadership. To be comprehensive in our investigation, we also incorporate two traditional innovation research variables, that is, expertise and involvement.

2.1. Opinion leadership
In a social system, communication flows from the source to the opinion leaders, who pass it on to the others in the social system (Katz & Lazarsfeld, 1955). Indeed, opinion leadership is considered an important factor in word-of-mouth communication, which contributes towards the success of an innovation (cf., [Childers, 1986] and [King and Summers, 1970]). Opinion leadership reflects an individual’s ability to influence other individuals’ attitudes or overt behavior in a desired way in a particular domain.
The important gate-keeping role of opinion leaders makes it imperative to understand its antecedents. We conjecture that one unexplored antecedent of opinion leadership is the social-identity function. People for whom a product performs the social-identity function feel that their desired social identity is contingent on owning the product. Product ownership and experience facilitates consumers learning about the product (Hoch & Deighton, 1989). Public access to ownership and, thereby, learning, paints the owner as informed. Therefore, we hypothesize that these owners are more likely to become opinion leaders.
H1: The social-identity function of attitudes positively impacts opinion leadership.

2.2. Innovativeness
Innovativeness is the bottom-line type of behavior in the diffusion process (Rogers, 1995). Innovators launch a new idea into a system by importing the idea from outside the system. Thus, it is important to develop an understanding of innovativeness. In addition, a good understanding of the process of diffusion of innovations requires an understanding of innovativeness (Gatignon & Robertson, 1991). Consistent with Gatignon and Robertson (1985), we contend that innovativeness is domain specific. For consumer products, this implies that a person might be an innovator in one product category (say computers) and a follower in another product category (say cars). Keeping this domain specific nature of innovativeness in mind, we adopt Goldsmith and Hofackers’ (1991) theorizing of innovativeness as the predisposition to learn about and adopt products in a specific domain.
Attitudes serving the social-identity function facilitate the acceptance of a person in a particular social setting (Katz, 1960). The attitudes, operationalized by the object, facilitate the image the person wants to portray. As the social-identity function facilitates the acceptance of a person in a particular social setting, the person is likely to adopt innovations in that domain. Thus, we expect the social-identity function to positively influence innovativeness.
H2: The social-identity function of attitudes positively impacts innovativeness.
One way to become an opinion leader in a product domain is to consistently own contemporary models of that product (Rogers, 1995). Thus, we expect innovativeness to determine opinion leadership.
H3: Innovativeness will have a positive influence on opinion leadership.

2.3. Expertise
Consumer knowledge manifests itself in finely differentiated and hierarchically organized knowledge structures, with well-developed consumption rules, and firmly entrenched beliefs about product performance (Sujan, 1985). Consumer knowledge has two components: familiarity and expertise (Jacoby, Troutman, Kuss & Mazursky, 1986). A consumer acquires familiarity by virtue of product related experiences, whereas expertise is the consumer’s ability to perform product related tasks successfully (Alba & Hutchinson, 1987).
Previous research has shown that innovators are heavy users of the relevant product category (Rogers, 1995). Although expertise is different from familiarity gained by virtue of experience with the product (Alba & Hutchinson, 1987), these two are expected to correlate (Gatignon & Robertson, 1991). Further, as experts understand a product domain better than novices do, the perceived risk of adoption of an innovation should be lower for experts. Thus, we expect expertise to be an antecedent of innovativeness.
H4: Expertise would have a positive influence on innovativeness.
Experts by definition are knowledgeable in the concerned domain. They own information not possessed but sought by others. In addition, the leadership status of opinion leaders is due to their technical competence (Rogers, 1995). Therefore, there is a higher likelihood for an expert to be an opinion leader in comparison to a novice.
H5: Expertise will have a positive influence on opinion leadership.

2.4. Involvement
Involvement signifies long-term interest in a domain and plays a central role in defining self-concept (Bloch, 1981). Consumer involvement with products depends on the personal relevance of that product [Celsi and Olson, 1988] and [Park and Hastak, 1994]. Persons involved with a product category gain experience with the product and acquire product-related knowledge. In this way, one can argue that involvement is an antecedent of expertise. Further, scholars assert that knowledge renders the world more comprehensible and predictable (Alba & Hutchinson, 1987). As this assertion is more cognitive than motivational, scholars have argued that knowledge is an unlikely candidate for arousing intense forms of personal involvement (Thomsen, Borgida & Lavine, 1995). To test these contradictory explanations, we propose two alternative hypotheses.
H6: Involvement will have a positive influence expertise.
H6ALT: Involvement does not influence expertise.
People are involved with a product category that is central to their beliefs about self [Bloch, 1981], [Boninger et al., 1995] and [Petty and Cacioppo, 1979]. People for whom the concerned product performs the social-identity function hold the product as being central to them and, therefore, are likely to be involved with the product.
H7: The social-identity function of attitudes is positively related to involvement.
Innovativeness in the concerned domain implies adopting relevant innovations, which requires an understanding of the innovation. One way to gain this understanding is by being involved with the concerned domain.
H8: Involvement positively influences innovativeness.
Finally, we expect involvement to positively influence opinion leadership. Opinion leaders have knowledge that is useful to others. This repository of information that the opinion leaders accumulate is by virtue of being involved with the concerned domain.
H9: Involvement positively influences opinion leadership.

3. The study

3.1. Measures
We use self-reported measures to operationalize products serving the social-identity function of attitudes for two products, viz., cars and computers. Indeed, we use a mixture of object based (product categories – Shavitt, 1989b) and individual difference (self-report measure) approaches to measure the social-identity function of attitudes (also see Clary, Snyder, Ridge, Miene & Haugen, 1994). We adopt four items, pertinent to our conceptualization, from the scale developed by Traylor and Joseph (1984) to measure context specific involvement and Childers’ (1986) version of the product-specific opinion leadership scale developed by King and Summers (1970). Following Sujan (1985), we use ten multiple-choice questions to measure expertise. To measure domain specific innovativeness, we adapted the three-item scale developed by Goldsmith and Hofacker (1991). Research on reverse polarity (cf., Herche & Engelland, 1996) recommends against using negatively worded items; thus we deleted the three reverse polarity items from Goldsmith and Hofacker’s (1991) six-item scale to get the three-item scale.
Based on our definition of the construct of social-identity function of attitudes, we generated four items to capture the construct. To verify face validity of these items, two fourth year doctoral students, not concerned with this research, performed a card sorting exercise. These students were given definitions of each construct along with cards with one item printed on each card (thus, they had 14 cards each). The classification of the items was 100% correct. Finally, three of the four items were used to measure social identity function (see section on measure validation – Table 1). The first item captures the construct directly by asking the importance of respondents’ friends’ knowing that the respondent owns the product. The second item emphasizes the importance of owning the “latest style”. Note that the emphasis is on “latest style” and not on either technological advancement (that is, innovativeness) or inner gratification (that is, involvement). The third item captures the construct indirectly by asking whether product ownership determines the respondents’ likelihood of preference for a person.

3.2. Context
Data were collected for two product categories: cars and computers. The order in which the items for these products were presented was randomized to eliminate any order effects. Cars were chosen because they are publicly consumed products that witness frequent incremental enhancements (innovations). On the other hand computers represent products that can be described as being primarily privately consumed and have also seen a lot of innovations lately. Following Bearden and Etzel (1982), we would expect social-identity function to be more important for publicly consumed products and involvement to be important for privately consumed products. We use a pretest to classify cars and computers along the public–private continuum. We adopted the six-item–six-point scale of Bearden and Etzel (1982) for the pretest. The mean for cars and computers were 5.32 and 3.17, with 1 signifying private consumption and 6 depicting public consumption.

3.3. Sample
Two hundred and twenty-six undergraduate students at a large mid-western US university participated in the study for extra credit. Two hundred and twenty-four responses were usable. All the students were within one year of finishing their undergraduate education and most of them already had jobs. One could argue that student sample is not suitable for our context, that is, cars and computers. On the contrary, as we are more interested in basic psychological processes than generalization, a student sample is appropriate. In fact, for research like ours that focuses on process generalization as opposed to effect generalization, scholars recommended using student samples (cf., [Calder et al., 1981], [Calder et al., 1982] and [Mook, 1983]. For instance, Kardes (1996, p. 287) recommends student research participants for “basic research on causal mechanisms”. Moreover, in a case very similar to ours, where students may not be experienced with purchasing cars or computers, Petty and Cacioppo (1996) persuasively argue that the effects of experience on judgment can be modeled effectively in laboratory settings with student participants. Finally, in a comparison of laboratory and field studies, Locke (1986) found substantively equivalent results in studies relating to industrial-organizational psychology, organizational behavior, and human resource management. Therefore, we think that there is ample precedence and evidence for us to use college students to study consumer behavior relating to cars and computers.

3.4. Results
We adopted Gerbing and Anderson’s (1988) scale development paradigm to develop a scale for the social-identity function of attitudes. To test our proposed model, we use Anderson and Gerbing’s (1988) two-step approach for structural equation modeling. First, exploratory factor analysis on the social-identity function of attitudes items identified one factor with an eigenvalue greater than 1 for cars, but in the case of computers there were two factors with eigenvalues greater than 1. The first factor had an eigenvalue of 6.40 and the second factor had an eigenvalue of 1.07. Eigenvalues are a measure of variance explained (or analytic importance) by a factor. An eigenvalue of 1 implies that a factor explains variance equal to that of an item. As there was a large difference between the eigenvalues of the first and second factor and the second eigenvalue was very close to one. Although, there are numerous rules of thumb that one could use to determine the number of factors (we refer the readers to Nunnally and Bernstein (1994, pp. 482–483) for a review), usually a cutoff of 1 for eigenvalues is used to suggest the number of factors. However, this cutoff is just a rule of thumb. As we have large differences between the first and the second eigenvalues and the second eigenvalue is very close to 1, we used our judgment and theory to interpret social-identity function as a unidimensional construct. Note that in the case of cars we had only one factor with this eigenvalue rule of thumb. Therefore, we believe that the exploratory factor analysis results indicate that the construct of the object’s invoking the social-identity function of attitudes is unidimensional.
Subsequently, we used LISREL to: (1) obtain the factor loadings for the confirmatory factor analysis and (2) test our hypotheses by estimating a structural equation model. LISREL is a software that uses maximum likelihood estimation to analyze covariance structures to model relationships between multiple latent variables. Usually, a researchers uses multiple items (observed variables) to measure a latent variable. Specifically, we estimate a four-construct, viz., social-identity function, opinion leadership, innovativeness, and involvement confirmatory factor model. High-standardized residuals, high modification indices, and low factors loading were used to refine the scales. We used low factor loadings, high modification indices, and high normalized residuals to refine the measures. Specifically, we use a multi-group analysis. As each respondent answered questions on both cars and computers, a multi-group analysis is appropriate. Table 1 has the factor loadings for the two product domains. Overall, the factor loadings are satisfactory. For the two products, the overall fit statistics for the measurement model signal mixed results. For the two products, the χ2 statistic is significant (see Table 1: χ2=848.9, d.f.=168, p=0.00). This anomaly of χ2 statistic not agreeing with other measures of overall goodness of fit statistics is common for larger sample sizes (cf., [Bagozzi and Warshaw, 1990] and [Stump and Heide, 1996]). The NNFI (0.7672) and CFI (0.8138) values are below the recommended levels, whereas GFI (0.9037) and Standardized RMR (0.0537) meet the recommended cutoff levels. At the very least, these fit statistics signal that future research should further refine the measure for the social-identity function.

Discriminant validity was assessed my means of average variance extracted and the confidence interval around φ2 between these constructs (Table 2). With two exceptions (one for each product), the average variance extracted was greater than the recommended 0.5 cut off (Bagozzi & Yi, 1988) and in all cases it was greater than the appropriate squared structural link (φ). Discriminant validity was further established as none of the confidence intervals around the maximum likelihood estimate of φ2 (±2 standard error) contained

1 (Anderson & Gerbing, 1988).
Nomological validity of the construct of social-identity function has been assessed in two ways: By means of correlation coefficients and by estimating the path model in Fig. 1.Table 3displays the correlation coefficients and Table 4summarizes the parameter estimates for the path model (Fig. 1). First, we find support for the nomological validity of the social-identity function of attitudes as it is positively correlated with involvement, innovativeness, and opinion leadership (Table 3). However, these are some anomalies in the case of expertise. Expertise is significantly correlated with innovativeness and opinion leadership in the case of cars, whereas for computers, expertise is positively correlated with opinion leadership. Note that for both cars and computers, the correlation between expertise and involvement was not significant.

We used structural equation modeling (LISREL 8.12) to test the proposed path model with the covariance matrix as the input (Anderson & Gerbing, 1988). Manifest indicators were created for each variable by averaging the items of each scale (cf., [Kenny, 1979] and [Williams and Hazar, 1986]). For the unidimensional scales (all our scales were unidimensional), we used the reliability coefficient (Cronbach’s alpha) to calculate the factor loadings and measurement error for each manifest variable. The path from the latent variable to its manifest variable was set equal to the square root of the reliability of the measured variable (cf., Settoon, Bennett & Liden 1996). The error variance for each manifest indicator was set equal to one minus reliability times the appropriate variance from the covariance matrix. The only exception was the reliability of expertise, which was set equal to 0.9 (Anderson & Gerbing, 1982). Specifically, we estimate the system of following equations (we summarize these equations.

In the case of car (H1), the social-identity function is an antecedent of opinion leadership, however this assertion is not supported for computers. Indeed, the social-identity function determines innovativeness for both product categories (H2) and innovativeness determined opinion leadership for both cars and computers (H3). Although expertise determined innovativeness only for computer (H4), it does seem to be an antecedent to opinion leadership for both product categories (H5). The hypothesis linking social-identity function and involvement (H7) was supported for both the product categories, however, involvement did not determine innovativeness or expertise (H8 and H6). Do note that we were skeptical of involvement determining expertise, as expertise is a cognitive construct, whereas involvement is a motivational construct (H6ALT). The hypothesis concerning the link between involvement and opinion leadership receives support in the case of computer (H9). In summary, we find support for (1) H2, H3, and H7 for both the product categories, (2) H1 for cars, and (3) H4 and H9 for computers

4. Discussion

The present study shows that for publicly consumed products, attitudes that serve a social-identity function play an important role in consumer innovativeness and opinion leadership. Attitudes that serve the social-identity function facilitate interpersonal interaction by communicating consumers’ consumption-related values and goals to other consumers. Social identity function is particularly important for publicly consumed products because such products reflect consumers’ personal tastes and preferences to a much greater extent than privately consumed products.
In the case of cars, experts tend to become opinion leaders; however, the social-identity function determines both innovativeness and opinion leadership. This pattern of results would be expected for a publicly consumed product, as visibility for publicly consumed goods is high (cf., Bearden & Etzel, 1982). For computers, expertise is an antecedent of both innovativeness and opinion leadership. The social-identity function is an antecedent of innovativeness but not of opinion leadership. The diminished role of social-identity function for computers relative to cars can be attributed to the fact that computers are privately consumed products.

4.1. Limitations

We highlight four limitations of our research. First, we only study the influence of personal characteristics of an individual on opinion leadership and innovativeness. Future research should investigate personal characteristics with other aspects of adoption and diffusion of consumer innovations such as interpersonal influence, characteristics of the innovation, the marketing activity of the innovating firm, and competitive activity (Gatignon & Robertson, 1985). Second, we only examine the social-identity function of attitudes as opposed to other functions that attitudes could perform. Given the strong results we obtain with respect to the social-identity function, future research should investigate other attitude functions. Third, as hypothesized, we did find our results to be context dependent – future research should further investigate this issue by examining other contexts. Finally, our confirmatory factor analysis results concerning the properties of our measurement model have scope to be improved.

4.2. Implications

Innovation plays a crucial role in consumer preference, competition among firms, and industry performance. Consumers prefer automobiles to streetcars, planes to trains, and small portable computers to large bulky computers. In a classic survey of 700 firms (60% industrial, 20% consumer durables, 20% consumer non-durables), new product innovations accounted for 28% of the observed growth of organizations (Booz, Allen, and Hamilton, 1982). More recently, Wind, Mahajan and Bayeless (1990) found that 25% of firms’ current sales resulted from new product innovations introduced within the last three years of the survey. All products go through the stages of introduction, growth, maturity, and decline, and recent technological advances and intensifying competition continue to shorten the duration of each of the product life cycle stages (Urban & Hauser, 1993).

Although innovation is critical to the growth and the success of any organization, consumers differ in their receptiveness to new product innovations. Some consumer segments actively seek out new product innovations and communicate the benefits of these innovations to other consumers. Other segments are much slower to adopt new innovations. Some segments (e.g., the laggard segment) are so resistant to innovation that by they time they adopt a new product it has already been replaced by a newer innovation (Urban & Hauser, 1993). Our research suggests that measuring consumers’ attitude functions concerning different product categories affords firms with the opportunity to identify those consumer segments that are likely to be most sensitive and responsive to new product offerings. Segments of consumers that hold attitudes serving the social-identity function are most likely adopt early, provided that the new product innovation is likely to be publicly rather than privately consumed. Not only are these segments likely to increase early sales and profits, they provide the added benefit of word-of-mouth communications to other segments that could also stimulate growth. Later adopters look to early adopters for information and opinion leadership that can greatly influence their purchase decisions.

In theoretic terms our research adds to current understanding of the process of adoption and diffusion of consumer innovations and sheds light on new antecedents of innovativeness and opinion leadership. Such an explanation of both innovativeness and opinion leadership has important theoretical implications as we are explaining two behavioral variables with the help of an attitudinal construct. The influence of attitude on behavior is well documented in literature, however which attitude influences what behavior is an important question for both researchers and practitioners. In addition, identifying opinion leaders and innovators is important for organizations as these leaders give the innovation legitimacy and innovators introduce it to the social system.

www.gunadarma.ac.id

Tugas softskill prilaku konsumen ( internasional )

Fadjar yudha kusuma
3EA05
12210485